title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Efficient Sampling for Selecting Important Nodes in Random Network | We consider the problem of selecting important nodes in a random network,
where the nodes connect to each other randomly with certain transition
probabilities. The node importance is characterized by the stationary
probabilities of the corresponding nodes in a Markov chain defined over the
network, as in Google's PageRank. Unlike deterministic network, the transition
probabilities in random network are unknown but can be estimated by sampling.
Under a Bayesian learning framework, we apply the first-order Taylor expansion
and normal approximation to provide a computationally efficient posterior
approximation of the stationary probabilities. In order to maximize the
probability of correct selection, we propose a dynamic sampling procedure which
uses not only posterior means and variances of certain interaction parameters
between different nodes, but also the sensitivities of the stationary
probabilities with respect to each interaction parameter. Numerical experiment
results demonstrate the superiority of the proposed sampling procedure.
| stat |
Hidden Markov models as recurrent neural networks: An application to
Alzheimer's disease | Hidden Markov models (HMMs) are commonly used for disease progression
modeling when the true patient health state is not fully known. Since HMMs may
have multiple local optima, performance can be improved by incorporating
additional patient covariates to inform estimation. To allow for this, we
develop hidden Markov recurrent neural networks (HMRNNs), a special case of
recurrent neural networks with the same likelihood function as a corresponding
discrete-observation HMM. The HMRNN can be combined with any other predictive
neural networks that take patient information as input, with all parameters
estimated simultaneously via gradient descent. Using a dataset of Alzheimer's
disease patients, we demonstrate how combining the HMRNN with other predictive
neural networks improves disease forecasting performance and offers a novel
clinical interpretation compared with a standard HMM trained via
expectation-maximization.
| stat |
Bayesian Inference of Local Projections with Roughness Penalty Priors | A local projection is a statistical framework that accounts for the
relationship between an exogenous variable and an endogenous variable, measured
at different time points. Local projections are often applied in impulse
response analyses and direct forecasting. While local projections are becoming
increasingly popular because of their robustness to misspecification and their
flexibility, they are less statistically efficient than standard methods, such
as vector autoregression. In this study, we seek to improve the statistical
efficiency of local projections by developing a fully Bayesian approach that
can be used to estimate local projections using roughness penalty priors. By
incorporating such prior-induced smoothness, we can use information contained
in successive observations to enhance the statistical efficiency of an
inference. We apply the proposed approach to an analysis of monetary policy in
the United States, showing that the roughness penalty priors successfully
estimate the impulse response functions and improve the predictive accuracy of
local projections.
| stat |
Estimation and false discovery control for the analysis of environmental
mixtures | The analysis of environmental mixtures is of growing importance in
environmental epidemiology, and one of the key goals in such analyses is to
identify exposures and their interactions that are associated with adverse
health outcomes. Typical approaches utilize flexible regression models combined
with variable selection to identify important exposures and estimate a
potentially nonlinear relationship with the outcome of interest. Despite this
surge in interest, no approaches to date can identify exposures and
interactions while controlling any form of error rates with respect to exposure
selection. We propose two novel approaches to estimating the health effects of
environmental mixtures that simultaneously 1) Estimate and provide valid
inference for the overall mixture effect, and 2) identify important exposures
and interactions while controlling the false discovery rate. We show that this
can lead to substantial power gains to detect weak effects of environmental
exposures. We apply our approaches to a study of persistent organic pollutants
and find that our approach is able to identify more interactions than existing
approaches.
| stat |
Interpretable, similarity-driven multi-view embeddings from
high-dimensional biomedical data | Similarity-driven multi-view linear reconstruction (SiMLR) is an algorithm
that exploits inter-modality relationships to transform large scientific
datasets into smaller, more well-powered and interpretable low-dimensional
spaces. SiMLR contributes a novel objective function for identifying joint
signal, regularization based on sparse matrices representing prior
within-modality relationships and an implementation that permits application to
joint reduction of large data matrices, each of which may have millions of
entries. We demonstrate that SiMLR outperforms closely related methods on
supervised learning problems in simulation data, a multi-omics cancer survival
prediction dataset and multiple modality neuroimaging datasets. Taken together,
this collection of results shows that SiMLR may be applied with default
parameters to joint signal estimation from disparate modalities and may yield
practically useful results in a variety of application domains.
| stat |
Random Tessellation Forests | Space partitioning methods such as random forests and the Mondrian process
are powerful machine learning methods for multi-dimensional and relational
data, and are based on recursively cutting a domain. The flexibility of these
methods is often limited by the requirement that the cuts be axis aligned. The
Ostomachion process and the self-consistent binary space partitioning-tree
process were recently introduced as generalizations of the Mondrian process for
space partitioning with non-axis aligned cuts in the two dimensional plane.
Motivated by the need for a multi-dimensional partitioning tree with non-axis
aligned cuts, we propose the Random Tessellation Process (RTP), a framework
that includes the Mondrian process and the binary space partitioning-tree
process as special cases. We derive a sequential Monte Carlo algorithm for
inference, and provide random forest methods. Our process is self-consistent
and can relax axis-aligned constraints, allowing complex inter-dimensional
dependence to be captured. We present a simulation study, and analyse gene
expression data of brain tissue, showing improved accuracies over other
methods.
| stat |
Simple and Honest Confidence Intervals in Nonparametric Regression | We consider the problem of constructing honest confidence intervals (CIs) for
a scalar parameter of interest, such as the regression discontinuity parameter,
in nonparametric regression based on kernel or local polynomial estimators. To
ensure that our CIs are honest, we use critical values that take into account
the possible bias of the estimator upon which the CIs are based. We show that
this approach leads to CIs that are more efficient than conventional CIs that
achieve coverage by undersmoothing or subtracting an estimate of the bias. We
give sharp efficiency bounds of using different kernels, and derive the optimal
bandwidth for constructing honest CIs. We show that using the bandwidth that
minimizes the maximum mean-squared error results in CIs that are nearly
efficient and that in this case, the critical value depends only on the rate of
convergence. For the common case in which the rate of convergence is
$n^{-2/5}$, the appropriate critical value for 95% CIs is 2.18, rather than the
usual 1.96 critical value. We illustrate our results in a Monte Carlo analysis
and an empirical application.
| stat |
Generalization Bounds for Convolutional Neural Networks | Convolutional neural networks (CNNs) have achieved breakthrough performances
in a wide range of applications including image classification, semantic
segmentation, and object detection. Previous research on characterizing the
generalization ability of neural networks mostly focuses on fully connected
neural networks (FNNs), regarding CNNs as a special case of FNNs without taking
into account the special structure of convolutional layers. In this work, we
propose a tighter generalization bound for CNNs by exploiting the sparse and
permutation structure of its weight matrices. As the generalization bound
relies on the spectral norm of weight matrices, we further study spectral norms
of three commonly used convolution operations including standard convolution,
depthwise convolution, and pointwise convolution. Theoretical and experimental
results both demonstrate that our bounds for CNNs are tighter than existing
bounds.
| stat |
On Regret with Multiple Best Arms | We study a regret minimization problem with the existence of multiple
best/near-optimal arms in the multi-armed bandit setting. We consider the case
when the number of arms/actions is comparable or much larger than the time
horizon, and make no assumptions about the structure of the bandit instance.
Our goal is to design algorithms that can automatically adapt to the unknown
hardness of the problem, i.e., the number of best arms. Our setting captures
many modern applications of bandit algorithms where the action space is
enormous and the information about the underlying instance/structure is
unavailable. We first propose an adaptive algorithm that is agnostic to the
hardness level and theoretically derive its regret bound. We then prove a lower
bound for our problem setting, which indicates: (1) no algorithm can be minimax
optimal simultaneously over all hardness levels; and (2) our algorithm achieves
a rate function that is Pareto optimal. With additional knowledge of the
expected reward of the best arm, we propose another adaptive algorithm that is
minimax optimal, up to polylog factors, over all hardness levels. Experimental
results confirm our theoretical guarantees and show advantages of our
algorithms over the previous state-of-the-art.
| stat |
On Random Subsampling of Gaussian Process Regression: A Graphon-Based
Analysis | In this paper, we study random subsampling of Gaussian process regression,
one of the simplest approximation baselines, from a theoretical perspective.
Although subsampling discards a large part of training data, we show provable
guarantees on the accuracy of the predictive mean/variance and its
generalization ability. For analysis, we consider embedding kernel matrices
into graphons, which encapsulate the difference of the sample size and enables
us to evaluate the approximation and generalization errors in a unified manner.
The experimental results show that the subsampling approximation achieves a
better trade-off regarding accuracy and runtime than the Nystr\"{o}m and random
Fourier expansion methods.
| stat |
Using machine learning to correct model error in data assimilation and
forecast applications | The idea of using machine learning (ML) methods to reconstruct the dynamics
of a system is the topic of recent studies in the geosciences, in which the key
output is a surrogate model meant to emulate the dynamical model. In order to
treat sparse and noisy observations in a rigorous way, ML can be combined to
data assimilation (DA). This yields a class of iterative methods in which, at
each iteration a DA step assimilates the observations, and alternates with a ML
step to learn the underlying dynamics of the DA analysis. In this article, we
propose to use this method to correct the error of an existent, knowledge-based
model. In practice, the resulting surrogate model is an hybrid model between
the original (knowledge-based) model and the ML model. We demonstrate
numerically the feasibility of the method using a two-layer, two-dimensional
quasi-geostrophic channel model. Model error is introduced by the means of
perturbed parameters. The DA step is performed using the strong-constraint
4D-Var algorithm, while the ML step is performed using deep learning tools. The
ML models are able to learn a substantial part of the model error and the
resulting hybrid surrogate models produce better short- to mid-range forecasts.
Furthermore, using the hybrid surrogate models for DA yields a significantly
better analysis than using the original model.
| stat |
Selection of Regression Models under Linear Restrictions for Fixed and
Random Designs | Many important modeling tasks in linear regression, including variable
selection (in which slopes of some predictors are set equal to zero) and
simplified models based on sums or differences of predictors (in which slopes
of those predictors are set equal to each other, or the negative of each other,
respectively), can be viewed as being based on imposing linear restrictions on
regression parameters. In this paper, we discuss how such models can be
compared using information criteria designed to estimate predictive measures
like squared error and Kullback-Leibler (KL) discrepancy, in the presence of
either deterministic predictors (fixed-X) or random predictors (random-X). We
extend the justifications for existing fixed-X criteria Cp, FPE and AICc, and
random-X criteria Sp and RCp, to general linear restrictions. We further
propose and justify a KL-based criterion, RAICc, under random-X for variable
selection and general linear restrictions. We show in simulations that the use
of the KL-based criteria AICc and RAICc results in better predictive
performance and sparser solutions than the use of squared error-based criteria,
including cross-validation.
| stat |
Identification and Model Testing in Linear Structural Equation Models
using Auxiliary Variables | We developed a novel approach to identification and model testing in linear
structural equation models (SEMs) based on auxiliary variables (AVs), which
generalizes a widely-used family of methods known as instrumental variables.
The identification problem is concerned with the conditions under which causal
parameters can be uniquely estimated from an observational, non-causal
covariance matrix. In this paper, we provide an algorithm for the
identification of causal parameters in linear structural models that subsumes
previous state-of-the-art methods. In other words, our algorithm identifies
strictly more coefficients and models than methods previously known in the
literature. Our algorithm builds on a graph-theoretic characterization of
conditional independence relations between auxiliary and model variables, which
is developed in this paper. Further, we leverage this new characterization for
allowing identification when limited experimental data or new substantive
knowledge about the domain is available. Lastly, we develop a new procedure for
model testing using AVs.
| stat |
Score-Based Parameter Estimation for a Class of Continuous-Time State
Space Models | We consider the problem of parameter estimation for a class of
continuous-time state space models. In particular, we explore the case of a
partially observed diffusion, with data also arriving according to a diffusion
process. Based upon a standard identity of the score function, we consider two
particle filter based methodologies to estimate the score function. Both
methods rely on an online estimation algorithm for the score function of
$\mathcal{O}(N^2)$ cost, with $N\in\mathbb{N}$ the number of particles. The
first approach employs a simple Euler discretization and standard particle
smoothers and is of cost $\mathcal{O}(N^2 + N\Delta_l^{-1})$ per unit time,
where $\Delta_l=2^{-l}$, $l\in\mathbb{N}_0$, is the time-discretization step.
The second approach is new and based upon a novel diffusion bridge
construction. It yields a new backward type Feynman-Kac formula in
continuous-time for the score function and is presented along with a particle
method for its approximation. Considering a time-discretization, the cost is
$\mathcal{O}(N^2\Delta_l^{-1})$ per unit time. To improve computational costs,
we then consider multilevel methodologies for the score function. We illustrate
our parameter estimation method via stochastic gradient approaches in several
numerical examples.
| stat |
Dyadic Reciprocity as a Function of Covariates | Reciprocity in dyadic interactions is common and a topic of interest across
disciplines. In some cases, reciprocity may be expected to be more or less
prevalent among certain kinds of dyads. In response to interest among
researchers in estimating dyadic reciprocity as a function of covariates, this
paper proposes an extension to the multilevel Social Relations Model. The
outcome variable is assumed to be a binomial proportion, as is commonly
encountered in observational and archival research. The approach draws on
principles of multilevel modeling to implement random intercepts and slopes
that vary among dyads. The corresponding variance function permits the
computation of a dyadic reciprocity correlation. The modeling approach can
potentially be integrated with other statistical models in the field of social
network analysis.
| stat |
matrixdist: An R Package for Inhomogeneous Phase-Type Distributions | Inhomogeneous phase-type distributions (IPH) are a broad class of laws which
arise from the absorption times of Markov jump processes. In the
time-homogeneous particular case, we recover phase-type (PH) distributions. In
matrix notation, various functionals corresponding to their distributional
properties are explicitly available and succinctly described. As the number of
parameters increases, IPH distributions may converge weakly to any probability
measure on the positive real line, making them particularly attractive
candidates for statistical modelling purposes. Contrary to PH distributions,
the IPH class allows for a wide range of tail behaviours, which often leads to
adequate estimation with a moderate number of parameters. One of the main
difficulties in estimating PH and IPH distributions is their large number of
matrix parameters. This drawback is best handled through the
expectation-maximisation (EM) algorithm, exploiting the underlying and
unobserved Markov structure. The matrixdist package presents tools for IPH
distributions to efficiently evaluate functionals, simulate, and carry out
maximum likelihood estimation through a three-step EM algorithm. Aggregated and
right-censored data are supported by the fitting routines, and in particular,
one may estimate time-to-event data, histograms, or discretised theoretical
distributions.
| stat |
Estimation and Validation of Ratio-based Conditional Average Treatment
Effects Using Observational Data | While sample sizes in randomized clinical trials are large enough to estimate
the average treatment effect well, they are often insufficient for estimation
of treatment-covariate interactions critical to studying data-driven precision
medicine. Observational data from real world practice may play an important
role in alleviating this problem. One common approach in trials is to predict
the outcome of interest with separate regression models in each treatment arm,
and estimate the treatment effect based on the contrast of the predictions.
Unfortunately, this simple approach may induce spurious treatment-covariate
interaction in observational studies when the regression model is misspecified.
Motivated by the need of modeling the number of relapses in multiple sclerosis
patients, where the ratio of relapse rates is a natural choice of the treatment
effect, we propose to estimate the conditional average treatment effect (CATE)
as the ratio of expected potential outcomes, and derive a doubly robust
estimator of this CATE in a semiparametric model of treatment-covariate
interactions. We also provide a validation procedure to check the quality of
the estimator on an independent sample. We conduct simulations to demonstrate
the finite sample performance of the proposed methods, and illustrate their
advantages on real data by examining the treatment effect of dimethyl fumarate
compared to teriflunomide in multiple sclerosis patients.
| stat |
Covariance-based sample selection for heterogeneous data: Applications
to gene expression and autism risk gene detection | Risk for autism can be influenced by genetic mutations in hundreds of genes.
Based on findings showing that genes with highly correlated gene expressions
are functionally interrelated, "guilt by association" methods such as DAWN have
been developed to identify these autism risk genes. Previous research analyzes
the BrainSpan dataset, which contains gene expression of brain tissues from
varying regions and developmental periods. Since the spatiotemporal properties
of brain tissue is known to affect the gene expression's covariance, previous
research have focused only on a specific subset of samples to avoid the issue
of heterogeneity. This leads to a potential loss of power when detecting risk
genes. In this article, we develop a new method called COBS (COvariance-Based
sample Selection) to find a larger and more homogeneous subset of samples that
share the same population covariance matrix for the downstream DAWN analysis.
To demonstrate COBS's effectiveness, we utilize genetic risk scores from two
sequential data freezes obtained in 2014 and 2019. We show COBS improves DAWN's
ability to predict risk genes detected in the newer data freeze when utilizing
the risk scores of the older data freeze as input.
| stat |
Exact Inference for Disease Prevalence Based on a Test with Unknown
Specificity and Sensitivity | To make informative public policy decisions in battling the ongoing COVID-19
pandemic, it is important to know the disease prevalence in a population. There
are two intertwined difficulties in estimating this prevalence based on testing
results from a group of subjects. First, the test is prone to measurement error
with unknown sensitivity and specificity. Second, the prevalence tends to be
low at the initial stage of the pandemic and we may not be able to determine if
a positive test result is a false positive due to the imperfect specificity of
the test. The statistical inference based on large sample approximation or
conventional bootstrap may not be sufficiently reliable and yield confidence
intervals that do not cover the true prevalence at the nominal level. In this
paper, we have proposed a set of 95% confidence intervals, whose validity is
guaranteed and doesn't depend on the sample size in the unweighted setting. For
the weighted setting, the proposed inference is equivalent to a class of hybrid
bootstrap methods, whose performance is also more robust to the sample size
than those based on asymptotic approximations. The methods are used to
reanalyze data from a study investigating the antibody prevalence in Santa
Clara county, California, which was the motivating example of this research, in
addition to several other seroprevalence studies where authors had tried to
correct their estimates for test performance. Extensive simulation studies have
been conducted to examine the finite-sample performance of the proposed
confidence intervals.
| stat |
Functional Principal Component Analysis for Extrapolating Multi-stream
Longitudinal Data | The advance of modern sensor technologies enables collection of multi-stream
longitudinal data where multiple signals from different units are collected in
real-time. In this article, we present a non-parametric approach to predict the
evolution of multi-stream longitudinal data for an in-service unit through
borrowing strength from other historical units. Our approach first decomposes
each stream into a linear combination of eigenfunctions and their corresponding
functional principal component (FPC) scores. A Gaussian process prior for the
FPC scores is then established based on a functional semi-metric that measures
similarities between streams of historical units and the in-service unit.
Finally, an empirical Bayesian updating strategy is derived to update the
established prior using real-time stream data obtained from the in-service
unit. Experiments on synthetic and real world data show that the proposed
framework outperforms state-of-the-art approaches and can effectively account
for heterogeneity as well as achieve high predictive accuracy.
| stat |
Disentangling Disentanglement in Variational Autoencoders | We develop a generalisation of disentanglement in VAEs---decomposition of the
latent representation---characterising it as the fulfilment of two factors: a)
the latent encodings of the data having an appropriate level of overlap, and b)
the aggregate encoding of the data conforming to a desired structure,
represented through the prior. Decomposition permits disentanglement, i.e.
explicit independence between latents, as a special case, but also allows for a
much richer class of properties to be imposed on the learnt representation,
such as sparsity, clustering, independent subspaces, or even intricate
hierarchical dependency relationships. We show that the $\beta$-VAE varies from
the standard VAE predominantly in its control of latent overlap and that for
the standard choice of an isotropic Gaussian prior, its objective is invariant
to rotations of the latent representation. Viewed from the decomposition
perspective, breaking this invariance with simple manipulations of the prior
can yield better disentanglement with little or no detriment to
reconstructions. We further demonstrate how other choices of prior can assist
in producing different decompositions and introduce an alternative training
objective that allows the control of both decomposition factors in a principled
manner.
| stat |
Generalised Boosted Forests | This paper extends recent work on boosting random forests to model
non-Gaussian responses. Given an exponential family $\mathbb{E}[Y|X] =
g^{-1}(f(X))$ our goal is to obtain an estimate for $f$. We start with an
MLE-type estimate in the link space and then define generalised residuals from
it. We use these residuals and some corresponding weights to fit a base random
forest and then repeat the same to obtain a boost random forest. We call the
sum of these three estimators a \textit{generalised boosted forest}. We show
with simulated and real data that both the random forest steps reduces test-set
log-likelihood, which we treat as our primary metric. We also provide a
variance estimator, which we can obtain with the same computational cost as the
original estimate itself. Empirical experiments on real-world data and
simulations demonstrate that the methods can effectively reduce bias, and that
confidence interval coverage is conservative in the bulk of the covariate
distribution.
| stat |
The survival-incorporated median versus the median in the survivors or
in the always-survivors: What are we measuring? And why? | Many clinical studies evaluate the benefit of treatment based on both
survival and other clinical outcomes. In these clinical studies, there are
situations that the clinical outcomes are truncated by death, where subjects
die before their clinical outcome is measured. Treating outcomes as "missing"
or "censored" due to death can be misleading for treatment effect evaluation.
We show that if we use the median in the survivors or in the always-survivors
to summarize clinical outcomes, we may conclude a trade-off exists between the
probability of survival and good clinical outcomes, even in settings where both
the probability of survival and the probability of any good clinical outcome
are better for one treatment. Therefore, we advocate not always treating death
as a mechanism through which clinical outcomes are missing, but rather as part
of the outcome measure. To account for the survival status, we describe the
survival-incorporated median as an alternative summary measure for outcomes in
the presence of death. The survival-incorporated median is the threshold such
that 50% of the population is alive with an outcome above that threshold. We
use conceptual examples to show that the survival-incorporated median provides
a simple and useful summary measure to inform clinical practice.
| stat |
Adaptive Quantile Low-Rank Matrix Factorization | Low-rank matrix factorization (LRMF) has received much popularity owing to
its successful applications in both computer vision and data mining. By
assuming noise to come from a Gaussian, Laplace or mixture of Gaussian
distributions, significant efforts have been made on optimizing the (weighted)
$L_1$ or $L_2$-norm loss between an observed matrix and its bilinear
factorization. However, the type of noise distribution is generally unknown in
real applications and inappropriate assumptions will inevitably deteriorate the
behavior of LRMF. On the other hand, real data are often corrupted by skew
rather than symmetric noise. To tackle this problem, this paper presents a
novel LRMF model called AQ-LRMF by modeling noise with a mixture of asymmetric
Laplace distributions. An efficient algorithm based on the
expectation-maximization (EM) algorithm is also offered to estimate the
parameters involved in AQ-LRMF. The AQ-LRMF model possesses the advantage that
it can approximate noise well no matter whether the real noise is symmetric or
skew. The core idea of AQ-LRMF lies in solving a weighted $L_1$ problem with
weights being learned from data. The experiments conducted on synthetic and
real datasets show that AQ-LRMF outperforms several state-of-the-art
techniques. Furthermore, AQ-LRMF also has the superiority over the other
algorithms in terms of capturing local structural information contained in real
images.
| stat |
Degrees of Freedom and Model Selection for k-means Clustering | This paper investigates the model degrees of freedom in k-means clustering.
An extension of Stein's lemma provides an expression for the effective degrees
of freedom in the k-means model. Approximating the degrees of freedom in
practice requires simplifications of this expression, however empirical studies
evince the appropriateness of our proposed approach. The practical relevance of
this new degrees of freedom formulation for k-means is demonstrated through
model selection using the Bayesian Information Criterion. The reliability of
this method is validated through experiments on simulated data as well as on a
large collection of publicly available benchmark data sets from diverse
application areas. Comparisons with popular existing techniques indicate that
this approach is extremely competitive for selecting high quality clustering
solutions. Code to implement the proposed approach is available in the form of
an R package from https://github.com/DavidHofmeyr/edfkmeans.
| stat |
A new sample-based algorithms to compute the total sensitivity index | Variance-based sensitivity indices have established themselves as a reference
among practitioners of sensitivity analysis of model output. It is not unusual
to consider a variance-based sensitivity analysis as informative if it produces
at least the first order sensitivity indices S_j and the so-called total-effect
sensitivity indices T_j for all the uncertain factors of the mathematical model
under analysis. Computational economy is critical in sensitivity analysis. It
depends mostly upon the number of model evaluations needed to obtain stable
values of the estimates. While efficient estimation procedures independent from
the number of factors under analysis are available for the first order indices,
this is less the case for the total sensitivity indices. When estimating T_j,
one can either use a sample-based approach, whose computational cost depends
fromon the number of factors, or approaches based on meta-modelling/emulators,
e.g. based on Gaussian processes. The present work focuses on sample-based
estimation procedures for T_j and tries different avenues to achieve an
algorithmic improvement over the designs proposed in the existing best
practices. We conclude that some proposed sample-based improvements found in
the literature do not work as claimed, and that improving on the existing best
practice is indeed fraught with difficulties. We motivate our conclusions
introducing the concepts of explorativity and efficiency of the design.
| stat |
Comparison of Clinical Episode Outcomes between Bundled Payments for
Care Improvement (BPCI) Initiative Participants and Non-Participants | Objective: To evaluate differences in major outcomes between Bundled Payments
for Care Improvement (BPCI) participating providers and non-participating
providers for both Major Joint Replacement of the Lower Extremity (MJRLE) and
Acute Myocardial Infarction (AMI) episodes. Methods: A
difference-in-differences approach estimated the differential change in
outcomes for Medicare beneficiaries who had an MJRLE or AMI at a BPCI
participating hospital between the baseline (January 2011 through September
2013) and intervention (October 2013 through December 2016) periods and
beneficiaries with the same episode (MJRLE or AMI) at a matched comparison
hospital. Main Outcomes and Measures: Medicare payments, LOS, and readmissions
during the episode, which includes the anchor hospitalization and the 90-day
post discharge period. Results: Mean total Medicare payments for an MJRLE
episode and the 90-day post discharge period declined $444 more (p < 0.0001)
for Medicare beneficiaries with episodes initiated in a BPCI-participating
provider than for the beneficiaries in a comparison provider. This reduction
was mainly due to reduced institutional post-acute care (PAC) payments. Slight
reductions in carrier payments and LOS were estimated. Readmission rates were
not statistically different between the BPCI and the comparison populations.
These findings suggest that PAC use can be reduced without adverse effects on
recovery from MJRLE. The lack of statistically significant differences in
effects for AMI could be explained by a smaller sample size or more
heterogenous recovery paths in AMI. Conclusions: Our findings suggest that, as
currently designed, bundled payments can be effective in reducing payments for
MJRLE episodes of care, but not necessarily for AMI. Most savings came from the
declines in PAC. These findings are consistent with the results reported in the
BPCI model evaluation for CMS.
| stat |
Binary Classification from Positive Data with Skewed Confidence | Positive-confidence (Pconf) classification [Ishida et al., 2018] is a
promising weakly-supervised learning method which trains a binary classifier
only from positive data equipped with confidence. However, in practice, the
confidence may be skewed by bias arising in an annotation process. The Pconf
classifier cannot be properly learned with skewed confidence, and consequently,
the classification performance might be deteriorated. In this paper, we
introduce the parameterized model of the skewed confidence, and propose the
method for selecting the hyperparameter which cancels out the negative impact
of skewed confidence under the assumption that we have the misclassification
rate of positive samples as a prior knowledge. We demonstrate the effectiveness
of the proposed method through a synthetic experiment with simple linear models
and benchmark problems with neural network models. We also apply our method to
drivers' drowsiness prediction to show that it works well with a real-world
problem where confidence is obtained based on manual annotation.
| stat |
Robust Estimation under Linear Mixed Models: The Minimum Density Power
Divergence Approach | Many real-life data sets can be analyzed using Linear Mixed Models (LMMs).
Since these are ordinarily based on normality assumptions, under small
deviations from the model the inference can be highly unstable when the
associated parameters are estimated by classical methods. On the other hand,
the density power divergence (DPD) family, which measures the discrepancy
between two probability density functions, has been successfully used to build
robust estimators with high stability associated with minimal loss in
efficiency. Here, we develop the minimum DPD estimator (MDPDE) for independent
but non identically distributed observations in LMMs. We prove the theoretical
properties, including consistency and asymptotic normality. The influence
function and sensitivity measures are studied to explore the robustness
properties. As a data based choice of the MDPDE tuning parameter $\alpha$ is
very important, we propose two candidates as "optimal" choices, where
optimality is in the sense of choosing the strongest downweighting that is
necessary for the particular data set. We conduct a simulation study comparing
the proposed MDPDE, for different values of $\alpha$, with the S-estimators,
M-estimators and the classical maximum likelihood estimator, considering
different levels of contamination. Finally, we illustrate the performance of
our proposal on a real-data example.
| stat |
Sampling Algorithms, from Survey Sampling to Monte Carlo Methods:
Tutorial and Literature Review | This paper is a tutorial and literature review on sampling algorithms. We
have two main types of sampling in statistics. The first type is survey
sampling which draws samples from a set or population. The second type is
sampling from probability distribution where we have a probability density or
mass function. In this paper, we cover both types of sampling. First, we review
some required background on mean squared error, variance, bias, maximum
likelihood estimation, Bernoulli, Binomial, and Hypergeometric distributions,
the Horvitz-Thompson estimator, and the Markov property. Then, we explain the
theory of simple random sampling, bootstrapping, stratified sampling, and
cluster sampling. We also briefly introduce multistage sampling, network
sampling, and snowball sampling. Afterwards, we switch to sampling from
distribution. We explain sampling from cumulative distribution function, Monte
Carlo approximation, simple Monte Carlo methods, and Markov Chain Monte Carlo
(MCMC) methods. For simple Monte Carlo methods, whose iterations are
independent, we cover importance sampling and rejection sampling. For MCMC
methods, we cover Metropolis algorithm, Metropolis-Hastings algorithm, Gibbs
sampling, and slice sampling. Then, we explain the random walk behaviour of
Monte Carlo methods and more efficient Monte Carlo methods, including
Hamiltonian (or hybrid) Monte Carlo, Adler's overrelaxation, and ordered
overrelaxation. Finally, we summarize the characteristics, pros, and cons of
sampling methods compared to each other. This paper can be useful for different
fields of statistics, machine learning, reinforcement learning, and
computational physics.
| stat |
Very Short Term Time-Series Forecasting of Solar Irradiance Without
Exogenous Inputs | This paper compares different forecasting methods and models to predict
average values of solar irradiance with a sampling time of 15 min over a
prediction horizon of up to 3 h. The methods considered only require historic
solar irradiance values, the current time and geographical location, i.e., no
exogenous inputs are used. Nearest neighbor regression (NNR) and autoregressive
integrated moving average (ARIMA) models are tested using different
hyperparameters, e.g., the number of lags, or the size of the training data
set, and data from different locations and seasons. The hyperparameters and
their effect on the forecast quality are analyzed to identify properties which
are likely to lead to good forecasts. Using these properties, a reduced search
space is derived to identify good forecasting models much faster.
| stat |
Hierarchical space-time modeling of exceedances with an application to
rainfall data | The statistical modeling of space-time extremes in environmental applications
is key to understanding complex dependence structures in original event data
and to generating realistic scenarios for impact models. In this context of
high-dimensional data, we propose a novel hierarchical model for high threshold
exceedances defined over continuous space and time by embedding a space-time
Gamma process convolution for the rate of an exponential variable, leading to
asymptotic independence in space and time. Its physically motivated anisotropic
dependence structure is based on geometric objects moving through space-time
according to a velocity vector. We demonstrate that inference based on weighted
pairwise likelihood is fast and accurate. The usefulness of our model is
illustrated by an application to hourly precipitation data from a study region
in Southern France, where it clearly improves on an alternative censored
Gaussian space-time random field model. While classical limit models based on
threshold-stability fail to appropriately capture relatively fast joint tail
decay rates between asymptotic dependence and classical independence, strong
empirical evidence from our application and other recent case studies motivates
the use of more realistic asymptotic independence models such as ours.
| stat |
Chance-Constrained Active Inference | Active Inference (ActInf) is an emerging theory that explains perception and
action in biological agents, in terms of minimizing a free energy bound on
Bayesian surprise. Goal-directed behavior is elicited by introducing prior
beliefs on the underlying generative model. In contrast to prior beliefs, which
constrain all realizations of a random variable, we propose an alternative
approach through chance constraints, which allow for a (typically small)
probability of constraint violation, and demonstrate how such constraints can
be used as intrinsic drivers for goal-directed behavior in ActInf. We
illustrate how chance-constrained ActInf weights all imposed (prior)
constraints on the generative model, allowing e.g., for a trade-off between
robust control and empirical chance constraint violation. Secondly, we
interpret the proposed solution within a message passing framework.
Interestingly, the message passing interpretation is not only relevant to the
context of ActInf, but also provides a general purpose approach that can
account for chance constraints on graphical models. The chance constraint
message updates can then be readily combined with other pre-derived message
update rules, without the need for custom derivations. The proposed
chance-constrained message passing framework thus accelerates the search for
workable models in general, and can be used to complement message-passing
formulations on generative neural models.
| stat |
A Simple and Fast Algorithm for L1-norm Kernel PCA | We present an algorithm for L1-norm kernel PCA and provide a convergence
analysis for it. While an optimal solution of L2-norm kernel PCA can be
obtained through matrix decomposition, finding that of L1-norm kernel PCA is
not trivial due to its non-convexity and non-smoothness. We provide a novel
reformulation through which an equivalent, geometrically interpretable problem
is obtained. Based on the geometric interpretation of the reformulated problem,
we present a fixed-point type algorithm that iteratively computes a binary
weight for each observation. As the algorithm requires only inner products of
data vectors, it is computationally efficient and the kernel trick is
applicable. In the convergence analysis, we show that the algorithm converges
to a local optimal solution in a finite number of steps. Moreover, we provide a
rate of convergence analysis, which has been never done for any L1-norm PCA
algorithm, proving that the sequence of objective values converges at a linear
rate. In numerical experiments, we show that the algorithm is robust in the
presence of entry-wise perturbations and computationally scalable, especially
in a large-scale setting. Lastly, we introduce an application to outlier
detection where the model based on the proposed algorithm outperforms the
benchmark algorithms.
| stat |
Bayesian model averaging for mortality forecasting using
leave-future-out validation | Predicting the evolution of mortality rates plays a central role for life
insurance and pension funds. Various stochastic frameworks have been developed
to model mortality patterns taking into account the main stylized facts driving
these patterns. However, relying on the prediction of one specific model can be
too restrictive and lead to some well documented drawbacks including model
misspecification, parameter uncertainty and overfitting. To address these
issues we first consider mortality modelling in a Bayesian Negative-Binomial
framework to account for overdispersion and the uncertainty about the parameter
estimates in a natural and coherent way. Model averaging techniques, which
consists in combining the predictions of several models, are then considered as
a response to model misspecifications. In this paper, we propose two methods
based on leave-future-out validation which are compared to the standard
Bayesian model averaging (BMA) based on marginal likelihood. Using
out-of-sample errors is a well-known workaround for overfitting issues. We show
that it also produces better forecasts. An intensive numerical study is carried
out over a large range of simulation setups to compare the performances of the
proposed methodologies. An illustration is then proposed on real-life mortality
datasets which includes a sensitivity analysis to a Covid-type scenario.
Overall, we found that both methods based on out-of-sample criterion outperform
the standard BMA approach in terms of prediction performance and robustness.
| stat |
Expert-Augmented Machine Learning | Machine Learning is proving invaluable across disciplines. However, its
success is often limited by the quality and quantity of available data, while
its adoption by the level of trust that models afford users. Human vs. machine
performance is commonly compared empirically to decide whether a certain task
should be performed by a computer or an expert. In reality, the optimal
learning strategy may involve combining the complementary strengths of man and
machine. Here we present Expert-Augmented Machine Learning (EAML), an automated
method that guides the extraction of expert knowledge and its integration into
machine-learned models. We use a large dataset of intensive care patient data
to predict mortality and show that we can extract expert knowledge using an
online platform, help reveal hidden confounders, improve generalizability on a
different population and learn using less data. EAML presents a novel framework
for high performance and dependable machine learning in critical applications.
| stat |
Quantile Surfaces -- Generalizing Quantile Regression to Multivariate
Targets | In this article, we present a novel approach to multivariate probabilistic
forecasting. Our approach is based on an extension of single-output quantile
regression (QR) to multivariate-targets, called quantile surfaces (QS). QS uses
a simple yet compelling idea of indexing observations of a probabilistic
forecast through direction and vector length to estimate a central tendency. We
extend the single-output QR technique to multivariate probabilistic targets. QS
efficiently models dependencies in multivariate target variables and represents
probability distributions through discrete quantile levels. Therefore, we
present a novel two-stage process. In the first stage, we perform a
deterministic point forecast (i.e., central tendency estimation). Subsequently,
we model the prediction uncertainty using QS involving neural networks called
quantile surface regression neural networks (QSNN). Additionally, we introduce
new methods for efficient and straightforward evaluation of the reliability and
sharpness of the issued probabilistic QS predictions. We complement this by the
directional extension of the Continuous Ranked Probability Score (CRPS) score.
Finally, we evaluate our novel approach on synthetic data and two currently
researched real-world challenges in two different domains: First, probabilistic
forecasting for renewable energy power generation, second, short-term cyclists
trajectory forecasting for autonomously driving vehicles. Especially for the
latter, our empirical results show that even a simple one-layer QSNN
outperforms traditional parametric multivariate forecasting techniques, thus
improving the state-of-the-art performance.
| stat |
Detecting and Classifying Outliers in Big Functional Data | We propose two new outlier detection methods, for identifying and classifying
different types of outliers in (big) functional data sets. The proposed methods
are based on an existing method called Massive Unsupervised Outlier Detection
(MUOD). MUOD detects and classifies outliers by computing for each curve, three
indices, all based on the concept of linear regression and correlation, which
measure outlyingness in terms of shape, magnitude and amplitude, relative to
the other curves in the data. 'Semifast-MUOD', the first method, uses a sample
of the observations in computing the indices, while 'Fast-MUOD', the second
method, uses the point-wise or L1 median in computing the indices. The
classical boxplot is used to separate the indices of the outliers from those of
the typical observations. Performance evaluation of the proposed methods using
simulated data show significant improvements compared to MUOD, both in outlier
detection and computational time. We show that Fast-MUOD is especially well
suited to handling big and dense functional datasets with very small
computational time compared to other methods. Further comparisons with some
recent outlier detection methods for functional data also show superior or
comparable outlier detection accuracy of the proposed methods. We apply the
proposed methods on weather, population growth, and video data.
| stat |
Statistical Evaluation of Spectral Methods for Anomaly Detection in
Networks | Monitoring of networks for anomaly detection has attracted a lot of attention
in recent years especially with the rise of connected devices and social
networks. This is of importance as anomaly detection could span a wide range of
application, from detecting terrorist cells in counter-terrorism efforts to
phishing attacks in social network circles. For this reason, numerous
techniques for anomaly detection have been introduced. However, application of
these techniques to more complex network models is hindered by various
challenges such as the size of the network being investigated, how much apriori
information is needed, the size of the anomalous graph, among others. A recent
technique introduced by Miller et al, which relies on a spectral framework for
anomaly detection, has the potential to address many of these challenges. In
their discussion of the spectral framework, three algorithms were proposed that
relied on the eigenvalues and eigenvectors of the residual matrix of a binary
network. The authors demonstrated the ability to detect anomalous subgraphs
that were less than 1% of the network size. However, to date, there is little
work that has been done to evaluate the statistical performance of these
algorithms. This study investigates the statistical properties of the spectral
methods, specifically the Chi-square and L1 norm algorithm proposed by Miller.
We will analyze the performance of the algorithm using simulated networks and
also extend the method's application to count networks. Finally we will make
some methodological improvements and recommendations to both algorithms.
| stat |
Exploring Maximum Entropy Distributions with Evolutionary Algorithms | This paper shows how to evolve numerically the maximum entropy probability
distributions for a given set of constraints, which is a variational calculus
problem. An evolutionary algorithm can obtain approximations to some well-known
analytical results, but is even more flexible and can find distributions for
which a closed formula cannot be readily stated. The numerical approach handles
distributions over finite intervals. We show that there are two ways of
conducting the procedure: by direct optimization of the Lagrangian of the
constrained problem, or by optimizing the entropy among the subset of
distributions which fulfill the constraints. An incremental evolutionary
strategy easily obtains the uniform, the exponential, the Gaussian, the
log-normal, the Laplace, among other distributions, once the constrained
problem is solved with any of the two methods. Solutions for mixed ("chimera")
distributions can be also found. We explain why many of the distributions are
symmetrical and continuous, but some are not.
| stat |
Regularized Estimation of High-dimensional Factor-Augmented Vector
Autoregressive (FAVAR) Models | A factor-augmented vector autoregressive (FAVAR) model is defined by a VAR
equation that captures lead-lag correlations amongst a set of observed
variables $X$ and latent factors $F$, and a calibration equation that relates
another set of observed variables $Y$ with $F$ and $X$. The latter equation is
used to estimate the factors that are subsequently used in estimating the
parameters of the VAR system. The FAVAR model has become popular in applied
economic research, since it can summarize a large number of variables of
interest as a few factors through the calibration equation and subsequently
examine their influence on core variables of primary interest through the VAR
equation. However, there is increasing need for examining lead-lag
relationships between a large number of time series, while incorporating
information from another high-dimensional set of variables. Hence, in this
paper we investigate the FAVAR model under high-dimensional scaling. We
introduce an appropriate identification constraint for the model parameters,
which when incorporated into the formulated optimization problem yields
estimates with good statistical properties. Further, we address a number of
technical challenges introduced by the fact that estimates of the VAR system
model parameters are based on estimated rather than directly observed
quantities. The performance of the proposed estimators is evaluated on
synthetic data. Further, the model is applied to commodity prices and reveals
interesting and interpretable relationships between the prices and the factors
extracted from a set of global macroeconomic indicators.
| stat |
Model-based Bayesian inference of disease outbreak dynamics with
invertible neural networks | Mathematical models in epidemiology strive to describe the dynamics and
important characteristics of infectious diseases. Apart from their scientific
merit, these models are often used to inform political decisions and
interventional measures during an ongoing outbreak. Since high-fidelity models
are often quite complex and analytically intractable, their applicability to
real data depends on powerful estimation algorithms. Moreover, uncertainty
quantification in such models is far from trivial, and different types of
uncertainty are often confounded. With this work, we introduce a novel coupling
between epidemiological models and specialized neural network architectures.
This coupling results in a powerful Bayesian inference framework capable of
principled uncertainty quantification and efficient amortized inference once
the networks have been trained on simulations from an arbitrarily complex
model. We illustrate the utility of our framework by applying it to real
Covid-19 cases from entire Germany and German federal states.
| stat |
Galton-Watson process and bayesian inference: A turnkey method for the
viability study of small populations | 1 Sharp prediction of extinction times is needed in biodiversity monitoring
and conservation management. 2 The Galton-Watson process is a classical
stochastic model for describing population dynamics. Its evolution is like the
matrix population model where offspring numbers are random. Extinction
probability, extinction time, abundance are well known and given by explicit
formulas. In contrast with the deterministic model, it can be applied to small
populations. 3 Parameters of this model can be estimated through the Bayesian
inference framework. This enables to consider non-arbitrary scenarios. 4 We
show how coupling Bayesian inference with the Galton-Watson model provides
several features: i) a flexible modelling approach with easily understandable
parameters ii) compatibility with the classical matrix population model (Leslie
type model) iii) A non-computational approach which then leads to more
information with less computing iv) a non-arbitrary choice for scenarios,
parameters... It can be seen to go one step further than the classical matrix
population model for the viability problem. 5 To illustrate these features, we
provide analysis details for two examples whose one of which is a real life
example.
| stat |
Bayesian Sparse Mediation Analysis with Targeted Penalization of Natural
Indirect Effects | Causal mediation analysis aims to characterize an exposure's effect on an
outcome and quantify the indirect effect that acts through a given mediator or
a group of mediators of interest. With the increasing availability of
measurements on a large number of potential mediators, like the epigenome or
the microbiome, new statistical methods are needed to simultaneously
accommodate high-dimensional mediators while directly target penalization of
the natural indirect effect (NIE) for active mediator identification. Here, we
develop two novel prior models for identification of active mediators in
high-dimensional mediation analysis through penalizing NIEs in a Bayesian
paradigm. Both methods specify a joint prior distribution on the
exposure-mediator effect and mediator-outcome effect with either (a) a
four-component Gaussian mixture prior or (b) a product threshold Gaussian
prior. By jointly modeling the two parameters that contribute to the NIE, the
proposed methods enable penalization on their product in a targeted way.
Resultant inference can take into account the four-component composite
structure underlying the NIE. We show through simulations that the proposed
methods improve both selection and estimation accuracy compared to other
competing methods. We applied our methods for an in-depth analysis of two
ongoing epidemiologic studies: the Multi-Ethnic Study of Atherosclerosis (MESA)
and the LIFECODES birth cohort. The identified active mediators in both studies
reveal important biological pathways for understanding disease mechanisms.
| stat |
PersLay: A Neural Network Layer for Persistence Diagrams and New Graph
Topological Signatures | Persistence diagrams, the most common descriptors of Topological Data
Analysis, encode topological properties of data and have already proved pivotal
in many different applications of data science. However, since the (metric)
space of persistence diagrams is not Hilbert, they end up being difficult
inputs for most Machine Learning techniques. To address this concern, several
vectorization methods have been put forward that embed persistence diagrams
into either finite-dimensional Euclidean space or (implicit) infinite
dimensional Hilbert space with kernels. In this work, we focus on persistence
diagrams built on top of graphs. Relying on extended persistence theory and the
so-called heat kernel signature, we show how graphs can be encoded by
(extended) persistence diagrams in a provably stable way. We then propose a
general and versatile framework for learning vectorizations of persistence
diagrams, which encompasses most of the vectorization techniques used in the
literature. We finally showcase the experimental strength of our setup by
achieving competitive scores on classification tasks on real-life graph
datasets.
| stat |
Bayesian Estimation of Two-Part Joint Models for a Longitudinal
Semicontinuous Biomarker and a Terminal Event with R-INLA: Interests for
Cancer Clinical Trial Evaluation | Two-part joint models for a longitudinal semicontinuous biomarker and a
terminal event have been recently introduced based on frequentist estimation.
The biomarker distribution is decomposed into a probability of positive value
and the expected value among positive values. Shared random effects can
represent the association structure between the biomarker and the terminal
event. The computational burden increases compared to standard joint models
with a single regression model for the biomarker. In this context, the
frequentist estimation implemented in the R package frailtypack can be
challenging for complex models (i.e., large number of parameters and dimension
of the random effects). As an alternative, we propose a Bayesian estimation of
two-part joint models based on the Integrated Nested Laplace Approximation
(INLA) algorithm to alleviate the computational burden and fit more complex
models. Our simulation studies show that R-INLA reduces the computation time
substantially as well as the variability of the parameter estimates and
improves the model convergence compared to frailtypack. We contrast the
Bayesian and frequentist approaches in the analysis of two randomized cancer
clinical trials (GERCOR and PRIME studies), where R-INLA suggests a stronger
association between the biomarker and the risk of event. Moreover, the Bayesian
approach was able to characterize subgroups of patients associated with
different responses to treatment in the PRIME study while frailtypack had
convergence issues. Our study suggests that the Bayesian approach using R-INLA
algorithm enables broader applications of the two-part joint model to clinical
applications.
| stat |
The design and statistical aspects of VIETNARMS: a strategic
post-licensing trial of multiple oral direct acting antiviral Hepatitis C
treatment strategies in Vietnam | Background Achieving hepatitis C elimination is hampered by the costs of
treatment and the need to treat hard-to-reach populations. Treatment access
could be widened by shortening treatment, but limited research means it is
unclear which strategies could achieve sufficiently high cure rates to be
acceptable. We present the statistical aspects of a multi-arm trial designed to
test multiple strategies simultaneously with a monitoring mechanism to detect
and stop those with unacceptably low cure rates quickly. Methods The VIETNARMS
trial will factorially randomise patients to three randomisations. We will use
Bayesian monitoring at interim analyses to detect and stop recruitment into
unsuccessful strategies, defined as a >0.95 posterior probability of the true
cure rate being <90%. Here, we tested the operating characteristics of the
stopping guideline, planned the timing of the interim analyses and explored
power at the final analysis. Results A beta(4.5, 0.5) prior for the true cure
rate produces <0.05 probability of incorrectly stopping a group with true cure
rate >90%. Groups with very low cure rates (<60%) are very likely (>0.9
probability) to stop after ~25% patients are recruited. Groups with moderately
low cure rates (80%) are likely to stop (0.7 probability) before the end of
recruitment. Interim analyses 7, 10, 13 and 18 months after recruitment
commences provide good probabilities of stopping inferior groups. For an
overall true cure rate of 95%, power is >90% to detect non-inferiority in the
regimen and strategy comparisons using 5% and 10% margins respectively,
regardless of the control cure rate, and to detect a 5% absolute difference in
the ribavirin comparison. Conclusions The operating characteristics of the
stopping guideline are appropriate and interim analyses can be timed to detect
failing groups at various stages.
| stat |
Manifold Gradient Descent Solves Multi-Channel Sparse Blind
Deconvolution Provably and Efficiently | Multi-channel sparse blind deconvolution, or convolutional sparse coding,
refers to the problem of learning an unknown filter by observing its circulant
convolutions with multiple input signals that are sparse. This problem finds
numerous applications in signal processing, computer vision, and inverse
problems. However, it is challenging to learn the filter efficiently due to the
bilinear structure of the observations with the respect to the unknown filter
and inputs, as well as the sparsity constraint. In this paper, we propose a
novel approach based on nonconvex optimization over the sphere manifold by
minimizing a smooth surrogate of the sparsity-promoting loss function. It is
demonstrated that manifold gradient descent with random initializations will
provably recover the filter, up to scaling and shift ambiguity, as soon as the
number of observations is sufficiently large under an appropriate random data
model. Numerical experiments are provided to illustrate the performance of the
proposed method with comparisons to existing ones.
| stat |
Universally Quantized Neural Compression | A popular approach to learning encoders for lossy compression is to use
additive uniform noise during training as a differentiable approximation to
test-time quantization. We demonstrate that a uniform noise channel can also be
implemented at test time using universal quantization (Ziv, 1985). This allows
us to eliminate the mismatch between training and test phases while maintaining
a completely differentiable loss function. Implementing the uniform noise
channel is a special case of the more general problem of communicating a
sample, which we prove is computationally hard if we do not make assumptions
about its distribution. However, the uniform special case is efficient as well
as easy to implement and thus of great interest from a practical point of view.
Finally, we show that quantization can be obtained as a limiting case of a soft
quantizer applied to the uniform noise channel, bridging compression with and
without quantization.
| stat |
Transportation Interventions Reshaping NYC Commute: the Probabilistic
Simulation Framework Assessing the Impacts of Ridesharing and Manhattan
Congestion Surcharge | Understanding holistic impact of planned transportation solutions and
interventions on urban systems is challenged by their complexity but critical
for decision making. The cornerstone for such impact assessments is estimating
the transportation mode-shift resulting from the intervention. And while
transportation planning has well-established models for the mode-choice
assessment such as the nested multinomial logit model, an individual choice
simulation could be better suited for addressing the mode-shift allowing to
consistently account for individual preferences. In addition, no model
perfectly represents the reality while the available ground truth data on the
actual transportation choices needed to infer the model is often incomplete or
inconsistent. The present paper addresses those challenges by offering an
individual mode-choice and mode-shift simulation model and the Bayesian
inference framework. It accounts for uncertainties in the data as well as the
model estimate and translates them into uncertainties of the resulting
mode-shift and the impacts. The framework is evaluated on the two intervention
cases: introducing ride-sharing for-hire-vehicles in NYC as well the recent
introduction of the Manhattan Congestion Surcharge. Being successfully
evaluated on the cases above, the framework can be used for assessing
mode-shift and resulting economic, social and environmental implications for
any future urban transportation solutions and policies being considered by
decision-makers or transportation companies.
| stat |
A primer on statistically validated networks | In this contribution we discuss some approaches of network analysis providing
information about single links or single nodes with respect to a null
hypothesis taking into account the heterogeneity of the system empirically
observed. With this approach, a selection of nodes and links is feasible when
the null hypothesis is statistically rejected. We focus our discussion on
approaches using (i) the so-called disparity filter and (ii) statistically
validated network in bipartite networks. For both methods we discuss the
importance of using multiple hypothesis test correction. Specific applications
of statistically validated networks are discussed. We also discuss how
statistically validated networks can be used to (i) pre-process large sets of
data and (ii) detect cores of communities that are forming the most close-knit
and stable subsets of clusters of nodes present in a complex system.
| stat |
Adaptive Multi-View ICA: Estimation of noise levels for optimal
inference | We consider a multi-view learning problem known as group independent
component analysis (group ICA), where the goal is to recover shared independent
sources from many views. The statistical modeling of this problem requires to
take noise into account. When the model includes additive noise on the
observations, the likelihood is intractable. By contrast, we propose Adaptive
multiView ICA (AVICA), a noisy ICA model where each view is a linear mixture of
shared independent sources with additive noise on the sources. In this setting,
the likelihood has a tractable expression, which enables either direct
optimization of the log-likelihood using a quasi-Newton method, or generalized
EM. Importantly, we consider that the noise levels are also parameters that are
learned from the data. This enables sources estimation with a closed-form
Minimum Mean Squared Error (MMSE) estimator which weights each view according
to its relative noise level. On synthetic data, AVICA yields better sources
estimates than other group ICA methods thanks to its explicit MMSE estimator.
On real magnetoencephalograpy (MEG) data, we provide evidence that the
decomposition is less sensitive to sampling noise and that the noise variance
estimates are biologically plausible. Lastly, on functional magnetic resonance
imaging (fMRI) data, AVICA exhibits best performance in transferring
information across views.
| stat |
Online Robust and Adaptive Learning from Data Streams | In online learning from non-stationary data streams, it is both necessary to
learn robustly to outliers and to adapt to changes of underlying data
generating mechanism quickly. In this paper, we refer to the former nature of
online learning algorithms as robustness and the latter as adaptivity. There is
an obvious tradeoff between them. It is a fundamental issue to quantify and
evaluate the tradeoff because it provides important information on the data
generating mechanism. However, no previous work has considered the tradeoff
quantitatively. We propose a novel algorithm called the Stochastic
approximation-based Robustness-Adaptivity algorithm (SRA) to evaluate the
tradeoff. The key idea of SRA is to update parameters of distribution or
sufficient statistics with the biased stochastic approximation scheme, while
dropping data points with large values of the stochastic update. We address the
relation between two parameters, one of which is the step size of the
stochastic approximation, and the other is the threshold parameter of the norm
of the stochastic update. The former controls the adaptivity and the latter
does the robustness. We give a theoretical analysis for the non-asymptotic
convergence of SRA in the presence of outliers, which depends on both the step
size and the threshold parameter. Since SRA is formulated on the
majorization-minimization principle, it is a general algorithm including many
algorithms, such as the online EM algorithm and stochastic gradient descent.
Empirical experiments for both synthetic and real datasets demonstrated that
SRA was superior to previous methods.
| stat |
Population-aware Hierarchical Bayesian Domain Adaptation via
Multiple-component Invariant Learning | While machine learning is rapidly being developed and deployed in health
settings such as influenza prediction, there are critical challenges in using
data from one environment in another due to variability in features; even
within disease labels there can be differences (e.g. "fever" may mean something
different reported in a doctor's office versus in an online app). Moreover,
models are often built on passive, observational data which contain different
distributions of population subgroups (e.g. men or women). Thus, there are two
forms of instability between environments in this observational transport
problem. We first harness knowledge from health to conceptualize the underlying
causal structure of this problem in a health outcome prediction task. Based on
sources of stability in the model, we posit that for human-sourced data and
health prediction tasks we can combine environment and population information
in a novel population-aware hierarchical Bayesian domain adaptation framework
that harnesses multiple invariant components through population attributes when
needed. We study the conditions under which invariant learning fails, leading
to reliance on the environment-specific attributes. Experimental results for an
influenza prediction task on four datasets gathered from different contexts
show the model can improve prediction in the case of largely unlabelled target
data from a new environment and different constituent population, by harnessing
both environment and population invariant information. This work represents a
novel, principled way to address a critical challenge by blending domain
(health) knowledge and algorithmic innovation. The proposed approach will have
a significant impact in many social settings wherein who and where the data
comes from matters.
| stat |
Curve Fitting from Probabilistic Emissions and Applications to Dynamic
Item Response Theory | Item response theory (IRT) models are widely used in psychometrics and
educational measurement, being deployed in many high stakes tests such as the
GRE aptitude test. IRT has largely focused on estimation of a single latent
trait (e.g. ability) that remains static through the collection of item
responses. However, in contemporary settings where item responses are being
continuously collected, such as Massive Open Online Courses (MOOCs), interest
will naturally be on the dynamics of ability, thus complicating usage of
traditional IRT models. We propose DynAEsti, an augmentation of the traditional
IRT Expectation Maximization algorithm that allows ability to be a continuously
varying curve over time. In the process, we develop CurvFiFE, a novel
non-parametric continuous-time technique that handles the
curve-fitting/regression problem extended to address more general probabilistic
emissions (as opposed to simply noisy data points). Furthermore, to accomplish
this, we develop a novel technique called grafting, which can successfully
approximate distributions represented by graphical models when other popular
techniques like Loopy Belief Propogation (LBP) and Variational Inference (VI)
fail. The performance of DynAEsti is evaluated through simulation, where we
achieve results comparable to the optimal of what is observed in the static
ability scenario. Finally, DynAEsti is applied to a longitudinal performance
dataset (80-years of competitive golf at the 18-hole Masters Tournament) to
demonstrate its ability to recover key properties of human performance and the
heterogeneous characteristics of the different holes. Python code for CurvFiFE
and DynAEsti is publicly available at github.com/chausies/DynAEstiAndCurvFiFE.
This is the full version of our ICDM 2019 paper.
| stat |
Feedback Coding for Active Learning | The iterative selection of examples for labeling in active machine learning
is conceptually similar to feedback channel coding in information theory: in
both tasks, the objective is to seek a minimal sequence of actions to encode
information in the presence of noise. While this high-level overlap has been
previously noted, there remain open questions on how to best formulate active
learning as a communications system to leverage existing analysis and
algorithms in feedback coding. In this work, we formally identify and leverage
the structural commonalities between the two problems, including the
characterization of encoder and noisy channel components, to design a new
algorithm. Specifically, we develop an optimal transport-based feedback coding
scheme called Approximate Posterior Matching (APM) for the task of active
example selection and explore its application to Bayesian logistic regression,
a popular model in active learning. We evaluate APM on a variety of datasets
and demonstrate learning performance comparable to existing active learning
methods, at a reduced computational cost. These results demonstrate the
potential of directly deploying concepts from feedback channel coding to design
efficient active learning strategies.
| stat |
Computing Shapley Effects for Sensitivity Analysis | Shapley effects are attracting increasing attention as sensitivity measures.
When the value function is the conditional variance, they account for the
individual and higher order effects of a model input. They are also well
defined under model input dependence. However, one of the issues associated
with their use is computational cost. We present a new algorithm that offers
major improvements for the computation of Shapley effects, reducing
computational burden by several orders of magnitude (from $k!\cdot k$ to $2^k$,
where $k$ is the number of inputs) with respect to currently available
implementations. The algorithm works in the presence of input dependencies. The
algorithm also makes it possible to estimate all generalized (Shapley-Owen)
effects for interactions.
| stat |
Bayesian Estimation of Attribute and Identification Disclosure Risks in
Synthetic Data | The synthetic data approach to data confidentiality has been actively
researched on, and for the past decade or so, a good number of high quality
work on developing innovative synthesizers, creating appropriate utility
measures and risk measures, among others, have been published. Comparing to a
large volume of work on synthesizers development and utility measures creation,
measuring risks has overall received less attention. This paper focuses on the
detailed construction of some Bayesian methods proposed for estimating
disclosure risks in synthetic data. In the processes of presenting attribute
and identification disclosure risks evaluation methods, we highlight key steps,
emphasize Bayesian thinking, illustrate with real application examples, and
discuss challenges and future research directions. We hope to give the readers
a comprehensive view of the Bayesian estimation procedures, enable synthetic
data researchers and producers to use these procedures to evaluate disclosure
risks, and encourage more researchers to work in this important growing field.
| stat |
Coefficients of Determination for Mixed-Effects Models | In consistency with the law of total variance, the coefficient of
determination, also known as $R^2$, is well-defined for linear regression
models to measure the proportion of variation in a variable explained by a set
of predictors. Following the same law, we will show that it is natural to
extend such a measure for linear mixed models. However, the heteroscedasticity
of a generalized linear model challenges further extension. By measuring the
change along the variance function responding to different means, we propose to
define proper coefficients of determination for generalized linear mixed
models, measuring the proportion of variation in the dependent variable modeled
by fixed effects, random effects, or both. As in the case of generalized linear
models, our measures can be calculated for general quasi-models with mixed
effects, which are only modeled with known link and variance functions. When
Gaussian models are considered, they reduce to those measures defined for
linear mixed models on the basis of the law of total variance.
| stat |
Distributional Null Hypothesis Testing with the T distribution | Null Hypothesis Significance Testing (NHST) has long been central to the
scientific project, guiding theory development and supporting evidence-based
intervention and decision-making. Recent years, however, have seen growing
awareness of serious problems with NHST as it is typically used, and hence to
proposals to limit the use of NHST techniques, to abandon these techniques and
move to alternative statistical approaches, or even to ban the use of NHST
entirely. These proposals are premature, because the observed problems with
NHST all arise as a consequence of a contingent and in many cases incorrect
choice: that of NHST testing against point-form nulls. We show that testing
against distributional, rather than point-form, nulls is better motivated
mathematically and experimentally, and that the use of distributional nulls
addresses many problems with the standard point-form NHST approach. We also
show that use of distributional nulls allows a form of null hypothesis testing
that takes into account both the statistical significance of a given result and
the probability of replication of that result in a new experiment. Rather than
abandoning NHST, we should use the NHST approach in its more general form, with
distributional rather than point-form nulls.
| stat |
Storchastic: A Framework for General Stochastic Automatic
Differentiation | Modelers use automatic differentiation of computation graphs to implement
complex Deep Learning models without defining gradient computations. However,
modelers often use sampling methods to estimate intractable expectations such
as in Reinforcement Learning and Variational Inference. Current methods for
estimating gradients through these sampling steps are limited: They are either
only applicable to continuous random variables and differentiable functions, or
can only use simple but high variance score-function estimators.
To overcome these limitations, we introduce Storchastic, a new framework for
automatic differentiation of stochastic computation graphs. Storchastic allows
the modeler to choose from a wide variety of gradient estimation methods at
each sampling step, to optimally reduce the variance of the gradient estimates.
Furthermore, Storchastic is provably unbiased for estimation of any-order
gradients, and generalizes variance reduction techniques to higher-order
gradient estimates. Finally, we implement Storchastic as a PyTorch library.
| stat |
ivmodel: An R Package for Inference and Sensitivity Analysis of
Instrumental Variables Models with One Endogenous Variable | We present a comprehensive R software ivmodel for analyzing instrumental
variables with one endogenous variable. The package implements a general class
of estimators called k- class estimators and two confidence intervals that are
fully robust to weak instruments. The package also provides power formulas for
various test statistics in instrumental variables. Finally, the package
contains methods for sensitivity analysis to examine the sensitivity of the
inference to instrumental variables assumptions. We demonstrate the software on
the data set from Card (1995), looking at the causal effect of levels of
education on log earnings where the instrument is proximity to a four-year
college.
| stat |
Accurate Characterization of Non-Uniformly Sampled Time Series using
Stochastic Differential Equations | Non-uniform sampling arises when an experimenter does not have full control
over the sampling characteristics of the process under investigation. Moreover,
it is introduced intentionally in algorithms such as Bayesian optimization and
compressive sensing. We argue that Stochastic Differential Equations (SDEs) are
especially well-suited for characterizing second order moments of such time
series. We introduce new initial estimates for the numerical optimization of
the likelihood, based on incremental estimation and initialization from
autoregressive models. Furthermore, we introduce model truncation as a purely
data-driven method to reduce the order of the estimated model based on the SDE
likelihood. We show the increased accuracy achieved with the new estimator in
simulation experiments, covering all challenging circumstances that may be
encountered in characterizing a non-uniformly sampled time series. Finally, we
apply the new estimator to experimental rainfall variability data.
| stat |
An interval-valued GARCH model for range-measured return processes | Range-measured return contains more information than the traditional
scalar-valued return. In this paper, we propose to model the [low, high] price
range as a random interval and suggest an interval-valued GARCH (Int-GARCH)
model for the corresponding range-measured return process. Under the general
framework of random sets, the model properties are investigated. Parameters are
estimated by the maximum likelihood method, and the asymptotic properties are
established. Empirical application to stocks and financial indices data sets
suggests that our Int-GARCH model overall outperforms the traditional GARCH for
both in-sample estimation and out-of-sample prediction of volatility.
| stat |
A different approach for choosing a threshold in peaks over threshold | Abstract In Extreme Value methodology the choice of threshold plays an
important role in efficient modelling of observations exceeding the threshold.
The threshold must be chosen high enough to ensure an unbiased extreme value
index but choosing the threshold too high results in uncontrolled variances.
This paper investigates a generalized model that can assist in the choice of
optimal threshold values in the \gamma positive domain. A Bayesian approach is
considered by deriving a posterior distribution for the unknown generalized
parameter. Using the properties of the posterior distribution allows for a
method to choose an optimal threshold without visual inspection.
| stat |
Robust Grouped Variable Selection Using Distributionally Robust
Optimization | We propose a Distributionally Robust Optimization (DRO) formulation with a
Wasserstein-based uncertainty set for selecting grouped variables under
perturbations on the data for both linear regression and classification
problems. The resulting model offers robustness explanations for Grouped Least
Absolute Shrinkage and Selection Operator (GLASSO) algorithms and highlights
the connection between robustness and regularization. We prove probabilistic
bounds on the out-of-sample loss and the estimation bias, and establish the
grouping effect of our estimator, showing that coefficients in the same group
converge to the same value as the sample correlation between covariates
approaches 1. Based on this result, we propose to use the spectral clustering
algorithm with the Gaussian similarity function to perform grouping on the
predictors, which makes our approach applicable without knowing the grouping
structure a priori. We compare our approach to an array of alternatives and
provide extensive numerical results on both synthetic data and a real large
dataset of surgery-related medical records, showing that our formulation
produces an interpretable and parsimonious model that encourages sparsity at a
group level and is able to achieve better prediction and estimation performance
in the presence of outliers.
| stat |
Stochastic Gradient Descent on a Tree: an Adaptive and Robust Approach
to Stochastic Convex Optimization | Online minimization of an unknown convex function over the interval $[0,1]$
is considered under first-order stochastic bandit feedback, which returns a
random realization of the gradient of the function at each query point. Without
knowing the distribution of the random gradients, a learning algorithm
sequentially chooses query points with the objective of minimizing regret
defined as the expected cumulative loss of the function values at the query
points in excess to the minimum value of the function. An approach based on
devising a biased random walk on an infinite-depth binary tree constructed
through successive partitioning of the domain of the function is developed.
Each move of the random walk is guided by a sequential test based on confidence
bounds on the empirical mean constructed using the law of the iterated
logarithm. With no tuning parameters, this learning algorithm is robust to
heavy-tailed noise with infinite variance and adaptive to unknown function
characteristics (specifically, convex, strongly convex, and nonsmooth). It
achieves the corresponding optimal regret orders (up to a $\sqrt{\log T}$ or a
$\log\log T$ factor) in each class of functions and offers better or matching
regret orders than the classical stochastic gradient descent approach which
requires the knowledge of the function characteristics for tuning the sequence
of step-sizes.
| stat |
k-GANs: Ensemble of Generative Models with Semi-Discrete Optimal
Transport | Generative adversarial networks (GANs) are the state of the art in generative
modeling. Unfortunately, most GAN methods are susceptible to mode collapse,
meaning that they tend to capture only a subset of the modes of the true
distribution. A possible way of dealing with this problem is to use an ensemble
of GANs, where (ideally) each network models a single mode. In this paper, we
introduce a principled method for training an ensemble of GANs using
semi-discrete optimal transport theory. In our approach, each generative
network models the transportation map between a point mass (Dirac measure) and
the restriction of the data distribution on a tile of a Voronoi tessellation
that is defined by the location of the point masses. We iteratively train the
generative networks and the point masses until convergence. The resulting
k-GANs algorithm has strong theoretical connection with the k-medoids
algorithm. In our experiments, we show that our ensemble method consistently
outperforms baseline GANs.
| stat |
On the parameter estimation of ARMA(p,q) model by approximate Bayesian
computation | In this paper, the parameter estimation of ARMA(p,q) model is given by
approximate Bayesian computation algorithm. In order to improve the sampling
efficiency of the algorithm, approximate Bayesian computation should select as
many statistics as possible with parameter information in low dimension.
Firstly, we use the autocorrelation coefficient of the first p+q order sample
as the statistic and obtain an approximate Bayesian estimation of the AR
coefficient, transforming the ARMA(p,q) model into the MA(q) model. Considering
the first q order sample autocorrelation functions and sample variance as the
statistics, the approximate Bayesian estimation of MA coefficient and white
noise variances can be given. The method mentioned above is more accurate and
powerful than the maximum likelihood estimation, which is verified by the
numerical simulations and experiment study.
| stat |
High-dimensional generalized semiparametric model for longitudinal data | This paper considers the problem of estimation in the generalized
semiparametric model for longitudinal data when the number of parameters
diverges with the sample size. A penalization type of generalized estimating
equation method is proposed, while we use the regression spline to approximate
the nonparametric component. The proposed procedure involves the specification
of the posterior distribution of the random effects, which cannot be evaluated
in a closed-form. However, it is possible to approximate this posterior
distribution by producing random draws from the distribution using a Metropolis
algorithm. Under some regularity conditions, the resulting estimators enjoy the
oracle properties, under the high-dimensional regime. Simulation studies are
carried out to assess the performance of our proposed method, and two real data
sets are analyzed to illustrate the procedure.
| stat |
Learning Social Networks from Text Data using Covariate Information | Describing and characterizing the impact of historical figures can be
challenging, but unraveling their social structures perhaps even more so.
Historical social network analysis methods can help and may also illuminate
people who have been overlooked by historians but turn out to be influential
social connection points. Text data, such as biographies, can be a useful
source of information about the structure of historical social networks but can
also introduce challenges in identifying links. The Local Poisson Graphical
Lasso model leverages the number of co-mentions in the text to measure
relationships between people and uses a conditional independence structure to
model a social network. This structure will reduce the tendency to overstate
the relationship between "friends of friends", but given the historical high
frequency of common names, without additional distinguishing information, we
can still introduce incorrect links. In this work, we extend the Local Poisson
Graphical Lasso model with a (multiple) penalty structure that incorporates
covariates giving increased link probabilities to people with shared covariate
information. We propose both greedy and Bayesian approaches to estimate the
penalty parameters. We present results on data simulated with characteristics
of historical networks and show that this type of penalty structure can improve
network recovery as measured by precision and recall. We also illustrate the
approach on biographical data of individuals who lived in early modern Britain,
targeting the period from 1500 to 1575.
| stat |
Isometric Gaussian Process Latent Variable Model for Dissimilarity Data | We propose a fully generative model where the latent variable respects both
the distances and the topology of the modeled data. The model leverages the
Riemannian geometry of the generated manifold to endow the latent space with a
well-defined stochastic distance measure, which is modeled as Nakagami
distributions. These stochastic distances are sought to be as similar as
possible to observed distances along a neighborhood graph through a censoring
process. The model is inferred by variational inference and is therefore fully
generative. We demonstrate how the new model can encode invariances in the
learned manifolds.
| stat |
Data-Driven Open Set Fault Classification and Fault Size Estimation
Using Quantitative Fault Diagnosis Analysis | Data-driven fault classification is complicated by imbalanced training data
and unknown fault classes. Fault diagnosis of dynamic systems is done by
detecting changes in time-series data, for example residuals, caused by faults
or system degradation. Different fault classes can result in similar residual
outputs, especially for small faults which can be difficult to distinguish from
nominal system operation. Analyzing how easy it is to distinguish data from
different fault classes is crucial during the design process of a diagnosis
system to evaluate if classification performance requirements can be met. Here,
a data-driven model of different fault classes is used based on the
Kullback-Leibler divergence. This is used to develop a framework for
quantitative fault diagnosis performance analysis and open set fault
classification. A data-driven fault classification algorithm is proposed which
can handle unknown faults and also estimate the fault size using training data
from known fault scenarios. To illustrate the usefulness of the proposed
methods, data have been collected from an engine test bench to illustrate the
design process of a data-driven diagnosis system, including quantitative fault
diagnosis analysis and evaluation of the developed open set fault
classification algorithm.
| stat |
Statistical Foundation of Variational Bayes Neural Networks | Despite the popularism of Bayesian neural networks in recent years, its use
is somewhat limited in complex and big data situations due to the computational
cost associated with full posterior evaluations. Variational Bayes (VB)
provides a useful alternative to circumvent the computational cost and time
complexity associated with the generation of samples from the true posterior
using Markov Chain Monte Carlo (MCMC) techniques. The efficacy of the VB
methods is well established in machine learning literature. However, its
potential broader impact is hindered due to a lack of theoretical validity from
a statistical perspective. However there are few results which revolve around
the theoretical properties of VB, especially in non-parametric problems. In
this paper, we establish the fundamental result of posterior consistency for
the mean-field variational posterior (VP) for a feed-forward artificial neural
network model. The paper underlines the conditions needed to guarantee that the
VP concentrates around Hellinger neighborhoods of the true density function.
Additionally, the role of the scale parameter and its influence on the
convergence rates has also been discussed. The paper mainly relies on two
results (1) the rate at which the true posterior grows (2) the rate at which
the KL-distance between the posterior and variational posterior grows. The
theory provides a guideline of building prior distributions for Bayesian NN
models along with an assessment of accuracy of the corresponding VB
implementation.
| stat |
Simplified Kalman filter for online rating: one-fits-all approach | In this work, we deal with the problem of rating in sports, where the skills
of the players/teams are inferred from the observed outcomes of the games. Our
focus is on the online rating algorithms which estimate the skills after each
new game by exploiting the probabilistic models of the relationship between the
skills and the game outcome. We propose a Bayesian approach which may be seen
as an approximate Kalman filter and which is generic in the sense that it can
be used with any skills-outcome model and can be applied in the individual --
as well as in the group-sports. We show how the well-know algorithms (such as
the Elo, the Glicko, and the TrueSkill algorithms) may be seen as instances of
the one-fits-all approach we propose. In order to clarify the conditions under
which the gains of the Bayesian approach over the simpler solutions can
actually materialize, we critically compare the known and the new algorithms by
means of numerical examples using the synthetic as well as the empirical data.
| stat |
Decomposing feature-level variation with Covariate Gaussian Process
Latent Variable Models | The interpretation of complex high-dimensional data typically requires the
use of dimensionality reduction techniques to extract explanatory
low-dimensional representations. However, in many real-world problems these
representations may not be sufficient to aid interpretation on their own, and
it would be desirable to interpret the model in terms of the original features
themselves. Our goal is to characterise how feature-level variation depends on
latent low-dimensional representations, external covariates, and non-linear
interactions between the two. In this paper, we propose to achieve this through
a structured kernel decomposition in a hybrid Gaussian Process model which we
call the Covariate Gaussian Process Latent Variable Model (c-GPLVM). We
demonstrate the utility of our model on simulated examples and applications in
disease progression modelling from high-dimensional gene expression data in the
presence of additional phenotypes. In each setting we show how the c-GPLVM can
extract low-dimensional structures from high-dimensional data sets whilst
allowing a breakdown of feature-level variability that is not present in other
commonly used dimensionality reduction approaches.
| stat |
An Initial Exploration of Bayesian Model Calibration for Estimating the
Composition of Rocks and Soils on Mars | The Mars Curiosity rover carries an instrument, ChemCam, designed to measure
the composition of surface rocks and soil using laser-induced breakdown
spectroscopy (LIBS). The measured spectra from this instrument must be analyzed
to identify the component elements in the target sample, as well as their
relative proportions. This process, which we call disaggregation, is
complicated by so-called matrix effects, which describe nonlinear changes in
the relative heights of emission lines as an unknown function of composition
due to atomic interactions within the LIBS plasma. In this work we explore the
use of the plasma physics code ATOMIC, developed at Los Alamos National
Laboratory, for the disaggregation task. ATOMIC has recently been used to model
LIBS spectra and can robustly reproduce matrix effects from first principles.
The ability of ATOMIC to predict LIBS spectra presents an exciting opportunity
to perform disaggregation in a manner not yet tried in the LIBS community,
namely via Bayesian model calibration. However, using it directly to solve our
inverse problem is computationally intractable due to the large parameter space
and the computation time required to produce a single output. Therefore we also
explore the use of emulators as a fast solution for this analysis. We discuss a
proof of concept Gaussian process emulator for disaggregating two-element
compounds of sodium and copper. The training and test datasets were simulated
with ATOMIC using a Latin hypercube design. After testing the performance of
the emulator, we successfully recover the composition of 25 test spectra with
Bayesian model calibration.
| stat |
Bayesian Updating and Sequential Testing: Overcoming Inferential
Limitations of Screening Tests | Bayes' Theorem confers inherent limitations on the accuracy of screening
tests as a function of disease prevalence. We have shown in previous work that
a testing system can tolerate significant drops in prevalence, up until a
certain well-defined point known as the $prevalence$ $threshold$, below which
the reliability of a positive screening test drops precipitously. Herein, we
establish a mathematical model to determine whether sequential testing
overcomes the aforementioned Bayesian limitations and thus improves the
reliability of screening tests. We show that for a desired positive predictive
value of $\rho$ that approaches $k$, the number of positive test iterations
$n_i$ needed is: $ n_i =\lim_{\rho \to
k}\left\lceil\frac{ln\left[\frac{\rho(\phi-1)}{\phi(\rho-1)}\right]}{ln\left[\frac{a}{1-b}\right]}\right\rceil$
where $n_i$ = number of testing iterations necessary to achieve $\rho$, the
desired positive predictive value, a = sensitivity, b = specificity, $\phi$ =
disease prevalence and $k$ = constant. Based on the aforementioned derivation,
we provide reference tables for the number of test iterations needed to obtain
a $\rho(\phi)$ of 50, 75, 95 and 99$\%$ as a function of various levels of
sensitivity, specificity and disease prevalence.
| stat |
Spatiotemporal Covariance Estimation by Shifted Partial Tracing | We consider the problem of covariance estimation for replicated space-time
processes from the functional data analysis perspective. Due to the challenges
to computational and statistical efficiency posed by the dimensionality of the
problem, common paradigms in the space-time processes literature typically
adopt parametric models, invariances, and/or separability. Replicated outcomes
may allow one to do away with parametric specifications, but considerations of
statistical and computational efficiency often compel the use of separability,
even though the assumption may fail in practice. In this paper, we consider the
problem of non-parametric covariance estimation, under "local" departures from
separability. Specifically, we consider a setting where the underlying random
field's second order structure is nearly separable, in that it may fail to be
separable only locally (either due to noise contamination or due to the
presence of a non-separable short-range dependent signal component). That is,
the covariance is an additive perturbation of a separable component by a
non-separable but banded component. We introduce non-parametric estimators
hinging on the novel concept of shifted partial tracing, which is capable of
estimating the model computationally efficiently under dense observation. Due
to the denoising properties of shifted partial tracing, our methods are shown
to yield consistent estimators of the separable part of the covariance even
under noisy discrete observation, without the need for smoothing. Further to
deriving the convergence rates and limit theorems, we also show that the
implementation of our estimators, including for the purpose of prediction,
comes at no computational overhead relative to a separable model. Finally, we
demonstrate empirical performance and computational feasibility of our methods
in an extensive simulation study and on a real data set.
| stat |
Causal mediation analysis in presence of multiple mediators uncausally
related | Mediation analysis aims at disentangling the effects of a treatment on an
outcome through alternative causal mechanisms and has become a popular practice
in biomedical and social science applications. The causal framework based on
counterfactuals is currently the standard approach to mediation, with important
methodological advances introduced in the literature in the last decade,
especially for simple mediation, that is with one mediator at the time. Among a
variety of alternative approaches, K. Imai et al. showed theoretical results
and developed an R package to deal with simple mediation as well as with
multiple mediation involving multiple mediators conditionally independent given
the treatment and baseline covariates. This approach does not allow to consider
the often encountered situation in which an unobserved common cause induces a
spurious correlation between the mediators. In this context, which we refer to
as mediation with uncausally related mediators, we show that, under appropriate
hypothesis, the natural direct and joint indirect effects are
non-parametrically identifiable. Moreover, we adopt the quasi-Bayesian
algorithm developed by Imai et al. and propose a procedure based on the
simulation of counterfactual distributions to estimate not only the direct and
joint indirect effects but also the indirect effects through individual
mediators. We study the properties of the proposed estimators through
simulations. As an illustration, we apply our method on a real data set from a
large cohort to assess the effect of hormone replacement treatment on breast
cancer risk through three mediators, namely dense mammographic area, nondense
area and body mass index.
| stat |
GrAMME: Semi-Supervised Learning using Multi-layered Graph Attention
Models | Modern data analysis pipelines are becoming increasingly complex due to the
presence of multi-view information sources. While graphs are effective in
modeling complex relationships, in many scenarios a single graph is rarely
sufficient to succinctly represent all interactions, and hence multi-layered
graphs have become popular. Though this leads to richer representations,
extending solutions from the single-graph case is not straightforward.
Consequently, there is a strong need for novel solutions to solve classical
problems, such as node classification, in the multi-layered case. In this
paper, we consider the problem of semi-supervised learning with multi-layered
graphs. Though deep network embeddings, e.g. DeepWalk, are widely adopted for
community discovery, we argue that feature learning with random node
attributes, using graph neural networks, can be more effective. To this end, we
propose to use attention models for effective feature learning, and develop two
novel architectures, GrAMME-SG and GrAMME-Fusion, that exploit the inter-layer
dependencies for building multi-layered graph embeddings. Using empirical
studies on several benchmark datasets, we evaluate the proposed approaches and
demonstrate significant performance improvements in comparison to
state-of-the-art network embedding strategies. The results also show that using
simple random features is an effective choice, even in cases where explicit
node attributes are not available.
| stat |
Nearly Optimal Adaptive Procedure with Change Detection for
Piecewise-Stationary Bandit | Multi-armed bandit (MAB) is a class of online learning problems where a
learning agent aims to maximize its expected cumulative reward while repeatedly
selecting to pull arms with unknown reward distributions. We consider a
scenario where the reward distributions may change in a piecewise-stationary
fashion at unknown time steps. We show that by incorporating a simple
change-detection component with classic UCB algorithms to detect and adapt to
changes, our so-called M-UCB algorithm can achieve nearly optimal regret bound
on the order of $O(\sqrt{MKT\log T})$, where $T$ is the number of time steps,
$K$ is the number of arms, and $M$ is the number of stationary segments.
Comparison with the best available lower bound shows that our M-UCB is nearly
optimal in $T$ up to a logarithmic factor. We also compare M-UCB with the
state-of-the-art algorithms in numerical experiments using a public Yahoo!
dataset to demonstrate its superior performance.
| stat |
Bus Travel Time Prediction: A Lognormal Auto-Regressive (AR) Modeling
Approach | Providing real time information about the arrival time of the transit buses
has become inevitable in urban areas to make the system more user-friendly and
advantageous over various other transportation modes. However, accurate
prediction of arrival time of buses is still a challenging problem in
dynamically varying traffic conditions especially under heterogeneous traffic
condition without lane discipline. One broad approach researchers have adopted
over the years is to segment the entire bus route into segments and work with
these segment travel times as the data input (from GPS traces) for prediction.
This paper adopts this approach and proposes predictive modelling approaches
which fully exploit the temporal correlations in the bus GPS data.
Specifically, we propose two approaches: (a) classical time-series approach
employing a seasonal AR model (b)unconventional linear, non-stationary AR
approach. The second approach is a novel technique and exploits the notion of
partial correlation for learning from data. A detailed analysis of the marginal
distributions of the data from Indian conditions (used here), revealed a
predominantly log-normal behavior. This aspect was incorporated into the above
proposed predictive models and statistically optimal prediction schemes in the
lognormal sense are utilized for all predictions. Both the above temporal
predictive modeling approaches predict ahead in time at each segment
independently. For real-time bus travel time prediction, one however needs to
predict across multiple segments ahead in space. Towards a complete solution,
the study also proposes an intelligent procedure to perform (real-time)
multi-section ahead travel-time predictions based on either of the above
proposed temporal models. Results showed a clear improvement in prediction
accuracy using the proposed methods,
| stat |
Closed-form variance estimators for weighted and stratified
dose-response function estimators using generalized propensity score | Propensity score methods are widely used in observational studies for
evaluating marginal treatment effects. The generalized propensity score (GPS)
is an extension of the propensity score framework, historically developed in
the case of binary exposures, for use with quantitative or continuous
exposures. In this paper, we proposed variance esti-mators for treatment effect
estimators on continuous outcomes. Dose-response functions (DRF) were estimated
through weighting on the inverse of the GPS, or using stratification. Variance
estimators were evaluated using Monte Carlo simulations. Despite the use of
stabilized weights, the variability of the weighted estimator of the DRF was
particularly high, and none of the variance estimators (a bootstrap-based
estimator, a closed-form estimator especially developped to take into account
the estimation step of the GPS, and a sandwich estimator) were able to
adequately capture this variability, resulting in coverages below to the
nominal value, particularly when the proportion of the variation in the
quantitative exposure explained by the covariates was 1 large. The stratified
estimator was more stable, and variance estima-tors (a bootstrap-based
estimator, a pooled linearized estimator, and a pooled model-based estimator)
more efficient at capturing the empirical variability of the parameters of the
DRF. The pooled variance estimators tended to overestimate the variance,
whereas the bootstrap estimator, which intrinsically takes into account the
estimation step of the GPS, resulted in correct variance estimations and
coverage rates. These methods were applied to a real data set with the aim of
assessing the effect of maternal body mass index on newborn birth weight.
| stat |
Modelling Network Interference with Multi-valued Treatments: the Causal
Effect of Immigration Policy on Crime Rates | Policy evaluation studies, which intend to assess the effect of an
intervention, face some statistical challenges: in real-world settings
treatments are not randomly assigned and the analysis might be further
complicated by the presence of interference between units. Researchers have
started to develop novel methods that allow to manage spillover mechanisms in
observational studies; recent works focus primarily on binary treatments.
However, many policy evaluation studies deal with more complex interventions.
For instance, in political science, evaluating the impact of policies
implemented by administrative entities often implies a multivariate approach,
as a policy towards a specific issue operates at many different levels and can
be defined along a number of dimensions. In this work, we extend the
statistical framework about causal inference under network interference in
observational studies, allowing for a multi-valued individual treatment and an
interference structure shaped by a weighted network. The estimation strategy is
based on a joint multiple generalized propensity score and allows one to
estimate direct effects, controlling for both individual and network
covariates. We follow the proposed methodology to analyze the impact of the
national immigration policy on the crime rate. We define a multi-valued
characterization of political attitudes towards migrants and we assume that the
extent to which each country can be influenced by another country is modeled by
an appropriate indicator, summarizing their cultural and geographical
proximity. Results suggest that implementing a highly restrictive immigration
policy leads to an increase of the crime rate and the estimated effects is
larger if we take into account interference from other countries.
| stat |
Conceptualizing experimental controls using the potential outcomes
framework | The goal of a well-controlled study is to remove unwanted variation when
estimating the causal effect of the intervention of interest. Experiments
conducted in the basic sciences frequently achieve this goal using experimental
controls, such as "negative" and "positive" controls, which are measurements
designed to detect systematic sources of unwanted variation. Here, we introduce
clear, mathematically precise definitions of experimental controls using
potential outcomes. Our definitions provide a unifying statistical framework
for fundamental concepts of experimental design from the biological and other
basic sciences. These controls are defined in terms of whether assumptions are
being made about a specific treatment level, outcome, or contrast between
outcomes. We discuss experimental controls as tools for researchers to wield in
designing experiments and detecting potential design flaws, including using
controls to diagnose unintended factors that influence the outcome of interest,
assess measurement error, and identify important subpopulations. We believe
that experimental controls are powerful tools for reproducible research that
are possibly underutilized by statisticians, epidemiologists, and social
science researchers.
| stat |
Most Powerful Test Sequences with Early Stopping Options | Sequential likelihood ratio testing is found to be most powerful in
sequential studies with early stopping rules when grouped data come from the
one-parameter exponential family.
First, to obtain this elusive result, the probability measure of a group
sequential design is constructed with support for all possible outcome events,
as is useful for designing an experiment prior to having data. This
construction identifies impossible events that are not part of the support. The
overall probability distribution is dissected into stage specific components.
These components are sub-densities of interim test statistics first described
by Armitage, McPherson and Rowe (1969) that are commonly used to create
stopping boundaries given an $\alpha$-spending function and a set of interim
analysis times. Likelihood expressions conditional on reaching a stage are
given to connect pieces of the probability anatomy together.
The reduction of the support caused by the adoption of an early stopping rule
induces sequential truncation (not nesting) in the probability distributions of
possible events. Multiple testing induces mixtures on the adapted support. Even
asymptotic distributions of inferential statistics are mixtures of truncated
distributions. In contrast to the classical result on local asymptotic
normality (Le Cam 1960), statistics that are asymptotically normal without
stopping options have asymptotic distributions that are mixtures of truncated
normal distributions under local alternatives with stopping options; under
fixed alternatives, asymptotic distributions of test statistics are degenerate.
| stat |
Network-Assisted Estimation for Large-dimensional Factor Model with
Guaranteed Convergence Rate Improvement | Network structure is growing popular for capturing the intrinsic relationship
between large-scale variables. In the paper we propose to improve the
estimation accuracy for large-dimensional factor model when a network structure
between individuals is observed. To fully excavate the prior network
information, we construct two different penalties to regularize the factor
loadings and shrink the idiosyncratic errors. Closed-form solutions are
provided for the penalized optimization problems. Theoretical results
demonstrate that the modified estimators achieve faster convergence rates and
lower asymptotic mean squared errors when the underlying network structure
among individuals is correct. An interesting finding is that even if the priori
network is totally misleading, the proposed estimators perform no worse than
conventional state-of-art methods. Furthermore, to facilitate the practical
application, we propose a data-driven approach to select the tuning parameters,
which is computationally efficient. We also provide an empirical criterion to
determine the number of common factors. Simulation studies and application to
the S&P100 weekly return dataset convincingly illustrate the superiority and
adaptivity of the new approach.
| stat |
Bayesian optimization with local search | Global optimization finds applications in a wide range of real world
problems. The multi-start methods are a popular class of global optimization
techniques, which are based on the ideas of conducting local searches at
multiple starting points. In this work we propose a new multi-start algorithm
where the starting points are determined in a Bayesian optimization framework.
Specifically, the method can be understood as to construct a new function by
conducting local searches of the original objective function, where the new
function attains the same global optima as the original one. Bayesian
optimization is then applied to find the global optima of the new local search
defined function.
| stat |
Statistical Failure Mechanism Analysis of Earthquakes Revealing Time
Relationships | If we assume that earthquakes are chaotic, and influenced locally then chaos
theory suggests that there should be a temporal association between earthquakes
in a local region that should be revealed with statistical examination. To date
no strong relationship has been shown (refs not prediction). However,
earthquakes are basically failures of structured material systems, and when
multiple failure mechanisms are present, prediction of failure is strongly
inhibited without first separating the mechanisms. Here we show that by
separating earthquakes statistically, based on their central tensor moment
structure, along lines first suggested by a separation into mechanisms
according to depth of the earthquake, a strong indication of temporal
association appears. We show this in earthquakes above 200 Km along the pacific
ring of fire, with a positive association in time between earthquakes of the
same statistical type and a negative association in time between earthquakes of
different types. Whether this can reveal either useful mechanistic information
to seismologists, or can result in useful forecasts remains to be seen.
| stat |
Tracing Network Evolution Using the PARAFAC2 Model | Characterizing time-evolving networks is a challenging task, but it is
crucial for understanding the dynamic behavior of complex systems such as the
brain. For instance, how spatial networks of functional connectivity in the
brain evolve during a task is not well-understood. A traditional approach in
neuroimaging data analysis is to make simplifications through the assumption of
static spatial networks. In this paper, without assuming static networks in
time and/or space, we arrange the temporal data as a higher-order tensor and
use a tensor factorization model called PARAFAC2 to capture underlying patterns
(spatial networks) in time-evolving data and their evolution. Numerical
experiments on simulated data demonstrate that PARAFAC2 can successfully reveal
the underlying networks and their dynamics. We also show the promising
performance of the model in terms of tracing the evolution of task-related
functional connectivity in the brain through the analysis of functional
magnetic resonance imaging data.
| stat |
Polynomial Cost of Adaptation for X -Armed Bandits | In the context of stochastic continuum-armed bandits, we present an algorithm
that adapts to the unknown smoothness of the objective function. We exhibit and
compute a polynomial cost of adaptation to the H{\"o}lder regularity for regret
minimization. To do this, we first reconsider the recent lower bound of
Locatelli and Carpentier [20], and define and characterize admissible rate
functions. Our new algorithm matches any of these minimal rate functions. We
provide a finite-time analysis and a thorough discussion about asymptotic
optimality.
| stat |
Another Approximation of the First-Passage Time Densities for the
Ratcliff Diffusion Decision Model | We present a novel method for approximating the probability density function
(PDF) of the first-passage times in the Ratcliff diffusion decision model
(DDM). We implemented this approximation method in $\texttt{C++}$ using the
$\texttt{R}$ package $\texttt{Rcpp}$ to utilize the faster $\texttt{C++}$
language while maintaining the $\texttt{R}$ language interface. In addition to
our novel approximation method, we also compiled all known approximation
methods for the DDM density function (with fixed and variable drift rate),
including previously unused combinations of techniques found in the relevant
literature. We ported these approximation methods to $\texttt{C++}$ and
optimized them to run in this new language. Given an acceptable error tolerance
in the value of the PDF approximation, we benchmarked all of these
approximation methods to compare their speed against each other and also
against commonly used $\texttt{R}$ functions from the literature. The results
of these tests show that our novel approximation method is not only orders of
magnitude faster than the current standards, but it is also faster than all of
the other approximation methods available even after translation and
optimization to the faster $\texttt{C++}$ language. All of these approximation
methods are bundled in the $\texttt{fddm}$ package for the $\texttt{R}$
statistical computing language; this package is available via CRAN, and the
source code is available on GitHub.
| stat |
Online Boosting for Multilabel Ranking with Top-k Feedback | We present online boosting algorithms for multilabel ranking with top-k
feedback, where the learner only receives information about the top k items
from the ranking it provides. We propose a novel surrogate loss function and
unbiased estimator, allowing weak learners to update themselves with limited
information. Using these techniques we adapt full information multilabel
ranking algorithms (Jung and Tewari, 2018) to the top-k feedback setting and
provide theoretical performance bounds which closely match the bounds of their
full information counterparts, with the cost of increased sample complexity.
These theoretical results are further substantiated by our experiments, which
show a small gap in performance between the algorithms for the top-k feedback
setting and that for the full information setting across various datasets.
| stat |
Exact Semiparametric Inference and Model Selection for Load-Sharing
Systems | As a specific proportional hazard rates model, sequential order statistics
can be used to describe the lifetimes of load-sharing systems. Inference for
these systems needs to account for small sample sizes, which are prevalent in
reliability applications. By exploiting the probabilistic structure of
sequential order statistics, we derive exact finite sample inference procedures
to test for the load-sharing parameters and for the nonparametrically specified
baseline distribution, treating the respective other part as a nuisance
quantity. This improves upon previous approaches for the model, which either
assume a fully parametric specification or rely on asymptotic results.
Simulations show that the tests derived are able to detect deviations from the
null hypothesis at small sample sizes. Critical values for a prominent case are
tabulated.
| stat |
Bayesian Non-Parametric Detection Heterogeneity in Ecological Models | Detection heterogeneity is inherent to ecological data, arising from factors
such as varied terrain or weather conditions, inconsistent sampling effort, or
heterogeneity of individuals themselves. Incorporating additional covariates
into a statistical model is one approach for addressing heterogeneity, but is
no guarantee that any set of measurable covariates will adequately address the
heterogeneity, and the presence of unmodelled heterogeneity has been shown to
produce biases in the resulting inferences. Other approaches for addressing
heterogeneity include the use of random effects, or finite mixtures of
homogeneous subgroups. Here, we present a non-parametric approach for modelling
detection heterogeneity for use in a Bayesian hierarchical framework. We employ
a Dirichlet process mixture which allows a flexible number of population
subgroups without the need to pre-specify this number of subgroups as in a
finite mixture. We describe this non-parametric approach, then consider its use
for modelling detection heterogeneity in two common ecological motifs:
capture-recapture and occupancy modelling. For each, we consider a homogeneous
model, finite mixture models, and the non-parametric approach. We compare these
approaches using two simulation studies, and observe the non-parametric
approach as the most reliable method for addressing varying degrees of
heterogeneity. We also present two real-data examples, and compare the
inferences resulting from each modelling approach. Analyses are carried out
using the \texttt{nimble} package for \texttt{R}, which provides facilities for
Bayesian non-parametric models.
| stat |
Product risk assessment: a Bayesian network approach | Product risk assessment is the overall process of determining whether a
product, which could be anything from a type of washing machine to a type of
teddy bear, is judged safe for consumers to use. There are several methods used
for product risk assessment, including RAPEX, which is the primary method used
by regulators in the UK and EU. However, despite its widespread use, we
identify several limitations of RAPEX including a limited approach to handling
uncertainty and the inability to incorporate causal explanations for using and
interpreting test data. In contrast, Bayesian Networks (BNs) are a rigorous,
normative method for modelling uncertainty and causality which are already used
for risk assessment in domains such as medicine and finance, as well as
critical systems generally. This article proposes a BN model that provides an
improved systematic method for product risk assessment that resolves the
identified limitations with RAPEX. We use our proposed method to demonstrate
risk assessments for a teddy bear and a new uncertified kettle for which there
is no testing data and the number of product instances is unknown. We show
that, while we can replicate the results of the RAPEX method, the BN approach
is more powerful and flexible.
| stat |
Unpacking the Drop in COVID-19 Case Fatality Rates: A Study of National
and Florida Line-Level Data | Since the COVID-19 pandemic first reached the United States, the case
fatality rate has fallen precipitously. Several possible explanations have been
floated, including greater detection of mild cases due to expanded testing,
shifts in age distribution among the infected, lags between confirmed cases and
reported deaths, improvements in treatment, mutations in the virus, and
decreased viral load as a result of mask-wearing. Using both Florida line-level
data and recently released (but incomplete) national line level data from April
1, 2020 to November 1, 2020 on cases, hospitalizations, and deaths--each
stratified by age--we unpack the drop in case fatality rate (CFR). Under the
hypothesis that improvements in treatment efficacy should correspond to
decreases in hospitalization fatality rate (HFR), we find that improvements in
the national data do not always match the story told by Florida data. In the
national data, treatment improvements between the first wave and the second
wave appear substantial, but modest when compared to the drop in aggregate CFR.
By contrast, possibly due to constrained resources in a much larger second
peak, Florida data suggests comparatively little difference between the first
and second wave, with HFR slightly increasing in every age group. However, by
November 1st, both Florida and national data suggest significant decreases in
age-stratified HFR since April 1st. By accounting for several confounding
factors, our analysis shows how age-stratified HFR can provide a more realistic
picture of treatment improvements than CFR. One key limitation of our analysis
is that the national line-level data remains incomplete and plagued by
artifacts. Our analysis highlights the crucial role that this data can play but
also the pressing need for public, complete, and high-quality age-stratified
line-level data for both cases, hospitalizations, and deaths for all states.
| stat |
Conditional Precedence Orders for Stochastic Comparison of Random
Variables | Most of the stochastic orders for comparing random variables, considered in
the literature, are afflicted with two main drawbacks: (i) lack of connex
property and (ii) lack of consideration of any dependence structure between the
random variables. Both these drawbacks can be overcome at the cost of
transitivity with the stochastic precedence order, which may seem to be a good
choice in particular when only two random variables are under consideration, a
situation where the question of transitivity does not arise. In this paper, we
show that even under such favorable conditions, stochastic precedence order may
direct to misleading conclusion in certain situations and develop variations of
the order to address the phenomenon.
| stat |
Distributional Clustering: A distribution-preserving clustering method | One key use of k-means clustering is to identify cluster prototypes which can
serve as representative points for a dataset. However, a drawback of using
k-means cluster centers as representative points is that such points distort
the distribution of the underlying data. This can be highly disadvantageous in
problems where the representative points are subsequently used to gain insights
on the data distribution, as these points do not mimic the distribution of the
data. To this end, we propose a new clustering method called "distributional
clustering", which ensures cluster centers capture the distribution of the
underlying data. We first prove the asymptotic convergence of the proposed
cluster centers to the data generating distribution, then present an efficient
algorithm for computing these cluster centers in practice. Finally, we
demonstrate the effectiveness of distributional clustering on synthetic and
real datasets.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.