title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Optimizing SAR data processing and thresholding for forest change
detection: an application for early deforestation warnings on eastern
Amazonia | The present work proposes a prototype for an operational method for early
deforestation detection of cloudy tropical rainforests. The proposed
methodology makes use of Sentinel-1 SAR data processed into the Google Earth
Engine platform for flag the areas where the probability of recent
deforestation is high. The evaluation of the results over a region on the
Eastern Amazon basin showed that copolarized data (VV band) offers the best
results in terms of producer's accuracy (95,4% for a 5% significance, 88,9% for
1% significance), while crosspolarized data (VH band) offered excellent results
in terms of user's accuracy (86% for a 5% significance, 100% for 1%
significance).
| stat |
Strengthening the Case for a Bayesian Approach to Car-following Model
Calibration and Validation using Probabilistic Programming | Compute and memory constraints have historically prevented traffic simulation
software users from fully utilizing the predictive models underlying them. When
calibrating car-following models, particularly, accommodations have included 1)
using sensitivity analysis to limit the number of parameters to be calibrated,
and 2) identifying only one set of parameter values using data collected from
multiple car-following instances across multiple drivers. Shortcuts are further
motivated by insufficient data set sizes, for which a driver may have too few
instances to fully account for the variation in their driving behavior. In this
paper, we demonstrate that recent technological advances can enable
transportation researchers and engineers to overcome these constraints and
produce calibration results that 1) outperform industry standard approaches,
and 2) allow for a unique set of parameters to be estimated for each driver in
a data set, even given a small amount of data. We propose a novel calibration
procedure for car-following models based on Bayesian machine learning and
probabilistic programming, and apply it to real-world data from a naturalistic
driving study. We also discuss how this combination of mathematical and
software tools can offer additional benefits such as more informative model
validation and the incorporation of true-to-data uncertainty into simulation
traces.
| stat |
Inference for two-stage sampling designs with application to a panel for
urban policy | Two-stage sampling designs are commonly used for household and health
surveys. To produce reliable estimators with assorted confidence intervals,
some basic statistical properties like consistency and asymptotic normality of
the Horvitz-Thompson estimator are desirable, along with the consistency of
assorted variance estimators. These properties have been mainly studied for
single-stage sampling designs. In this work, we prove the consistency of the
Horvitz-Thompson estimator and of associated variance estimators for a general
class of two-stage sampling designs, under mild assumptions. We also study
two-stage sampling with a large entropy sampling design at the first stage, and
prove that the Horvitz-Thompson estimator is asymptotically normally
distributed through a coupling argument. When the first-stage sampling fraction
is negligible, simplified variance estimators which do not require estimating
the variance within the Primary Sampling Units are proposed, and shown to be
consistent. An application to a panel for urban policy, which is the initial
motivation for this work, is also presented.
| stat |
False Discovery Rates in Biological Networks | The increasing availability of data has generated unprecedented prospects for
network analyses in many biological fields, such as neuroscience (e.g., brain
networks), genomics (e.g., gene-gene interaction networks), and ecology (e.g.,
species interaction networks). A powerful statistical framework for estimating
such networks is Gaussian graphical models, but standard estimators for the
corresponding graphs are prone to large numbers of false discoveries. In this
paper, we introduce a novel graph estimator based on knockoffs that imitate the
partial correlation structures of unconnected nodes. We show that this new
estimator guarantees accurate control of the false discovery rate in theory,
simulations, and biological applications, and we provide easy-to-use R code.
| stat |
A dose-effect network meta-analysis model: an application in
antidepressants | Network meta-analysis (NMA) has been used to answer a range of clinical
questions about the preferable intervention for a given condition. Although the
effectiveness and safety of pharmacological agents depend on the dose
administered, NMA applications typically ignore the role that drugs dosage play
on the results. This leads to more heterogeneity in the network. In this paper
we present a suite of network meta-analysis models that incorporates the
dose-effect relationship (DE-NMA) using restricted cubic splines (RCS). We
extend the model into a dose-effect network meta-regression to account for
study-level covariates and for groups of agents in a class-effect DE-NMA model.
We apply the models to a network of aggregate data about the efficacy of 21
antidepressants and placebo for depression. We found that all antidepressants
are more efficacious than placebo after a certain dose. We also identify the
dose level in which each antidepressant effect exceeds that of placebo and
estimate the dose beyond the effect of the antidepressants no longer increases.
The DE-NMA model with RCS takes a flexible approach to modelling the
dose-effect relationship in multiple interventions, so decision-makers can use
them to inform treatment choice.
| stat |
A Survey of Complex-Valued Neural Networks | Artificial neural networks (ANNs) based machine learning models and
especially deep learning models have been widely applied in computer vision,
signal processing, wireless communications, and many other domains, where
complex numbers occur either naturally or by design. However, most of the
current implementations of ANNs and machine learning frameworks are using real
numbers rather than complex numbers. There are growing interests in building
ANNs using complex numbers, and exploring the potential advantages of the
so-called complex-valued neural networks (CVNNs) over their real-valued
counterparts. In this paper, we discuss the recent development of CVNNs by
performing a survey of the works on CVNNs in the literature. Specifically, a
detailed review of various CVNNs in terms of activation function, learning and
optimization, input and output representations, and their applications in tasks
such as signal processing and computer vision are provided, followed by a
discussion on some pertinent challenges and future research directions.
| stat |
Provably Efficient Reinforcement Learning with Aggregated States | We establish that an optimistic variant of Q-learning applied to a
fixed-horizon episodic Markov decision process with an aggregated state
representation incurs regret $\tilde{\mathcal{O}}(\sqrt{H^5 M K} + \epsilon
HK)$, where $H$ is the horizon, $M$ is the number of aggregate states, $K$ is
the number of episodes, and $\epsilon$ is the largest difference between any
pair of optimal state-action values associated with a common aggregate state.
Notably, this regret bound does not depend on the number of states or actions
and indicates that asymptotic per-period regret is no greater than $\epsilon$,
independent of horizon. To our knowledge, this is the first such result that
applies to reinforcement learning with nontrivial value function approximation
without any restrictions on transition probabilities.
| stat |
A regression-based method for detecting publication bias in multivariate
meta-analysis | Publication bias occurs when the publication of research results depends not
only on the quality of the research but also on its nature and direction. The
consequence is that published studies may not be truly representative of all
valid studies undertaken, and this bias may threaten the validity of systematic
reviews and meta-analyses - on which evidence-based medicine increasingly
relies. Multivariate meta-analysis has recently received increasing attention
for its ability reducing potential bias and improving statistical efficiency by
borrowing information across outcomes. However, detecting and accounting for
publication bias are more challenging in multivariate meta-analysis setting
because some studies may be completely unpublished whereas some studies may
selectively report part of multiple outcomes. In this paper, we propose a score
test for jointly testing publication bias for multiple outcomes, which is novel
to the multivariate setting. The proposed test is a natural multivariate
extension of the univariate Egger's test, and can handle the above mentioned
scenarios simultaneously, It accounts for correlations among multivariate
outcomes, while allowing different types of outcomes, and can borrow
information across outcomes. The proposed test is shown to be more powerful
than the Egger's test, Begg's test and Trim and Fill method through simulation
studies. Two data analyses are given to illustrate the performance of the
proposed test in practice.
| stat |
Efficient Stochastic Gradient Descent for Learning with Distributionally
Robust Optimization | Distributionally robust optimization (DRO) problems are increasingly seen as
a viable method to train machine learning models for improved model
generalization. These min-max formulations, however, are more difficult to
solve. We therefore provide a new stochastic gradient descent algorithm to
efficiently solve this DRO formulation. Our approach applies gradient descent
to the outer minimization formulation and estimates the gradient of the inner
maximization based on a sample average approximation. The latter uses a subset
of the data in each iteration, progressively increasing the subset size to
ensure convergence. Theoretical results include establishing the optimal manner
for growing the support size to balance a fundamental tradeoff between
stochastic error and computational effort. Empirical results demonstrate the
significant benefits of our approach over previous work, and also illustrate
how learning with DRO can improve generalization.
| stat |
Bayesian Modeling of Multiple Structural Connectivity Networks During
the Progression of Alzheimer's Disease | Alzheimer's disease is the most common neurodegenerative disease. The aim of
this study is to infer structural changes in brain connectivity resulting from
disease progression using cortical thickness measurements from a cohort of
participants who were either healthy control, or with mild cognitive
impairment, or Alzheimer's disease patients. For this purpose, we develop a
novel approach for inference of multiple networks with related edge values
across groups. Specifically, we infer a Gaussian graphical model for each group
within a joint framework, where we rely on Bayesian hierarchical priors to link
the precision matrix entries across groups. Our proposal differs from existing
approaches in that it flexibly learns which groups have the most similar edge
values, and accounts for the strength of connection (rather than only edge
presence or absence) when sharing information across groups. Our results
identify key alterations in structural connectivity which may reflect
disruptions to the healthy brain, such as decreased connectivity within the
occipital lobe with increasing disease severity. We also illustrate the
proposed method through simulations, where we demonstrate its performance in
structure learning and precision matrix estimation with respect to alternative
approaches.
| stat |
Geometric Scattering on Manifolds | The Euclidean scattering transform was introduced nearly a decade ago to
improve the mathematical understanding of the success of convolutional neural
networks (ConvNets) in image data analysis and other tasks. Inspired by recent
interest in geometric deep learning, which aims to generalize ConvNets to
manifold and graph-structured domains, we generalize the scattering transform
to compact manifolds. Similar to the Euclidean scattering transform, our
geometric scattering transform is based on a cascade of designed filters and
pointwise nonlinearities, which enables rigorous analysis of the feature
extraction provided by scattering layers. Our main focus here is on theoretical
understanding of this geometric scattering network, while setting aside
implementation aspects, although we remark that application of similar
transforms to graph data analysis has been studied recently in related work.
Our results establish conditions under which geometric scattering provides
localized isometry invariant descriptions of manifold signals, which are also
stable to families of diffeomorphisms formulated in intrinsic manifolds terms.
These results not only generalize the deformation stability and local
roto-translation invariance of Euclidean scattering, but also demonstrate the
importance of linking the used filter structures (e.g., in geometric deep
learning) to the underlying manifold geometry, or the data geometry it
represents.
| stat |
A numerically stable algorithm for integrating Bayesian models using
Markov melding | When statistical analyses consider multiple data sources, Markov melding
provides a method for combining the source-specific Bayesian models. Models
often contain different quantities of information due to variation in the
richness of model-specific data, or availability of model-specific prior
information. We show that this can make the multi-stage Markov chain Monte
Carlo sampler employed by Markov melding unstable and unreliable. We propose a
robust multi-stage algorithm that estimates the required prior marginal
self-density ratios using weighted samples, dramatically improving accuracy in
the tails of the distribution, thus stabilising the algorithm and providing
reliable inference. We demonstrate our approach using an evidence synthesis for
inferring HIV prevalence, and an evidence synthesis of A/H1N1 influenza.
| stat |
Machine Learning in Least-Squares Monte Carlo Proxy Modeling of Life
Insurance Companies | Under the Solvency II regime, life insurance companies are asked to derive
their solvency capital requirements from the full loss distributions over the
coming year. Since the industry is currently far from being endowed with
sufficient computational capacities to fully simulate these distributions, the
insurers have to rely on suitable approximation techniques such as the
least-squares Monte Carlo (LSMC) method. The key idea of LSMC is to run only a
few wisely selected simulations and to process their output further to obtain a
risk-dependent proxy function of the loss. In this paper, we present and
analyze various adaptive machine learning approaches that can take over the
proxy modeling task. The studied approaches range from ordinary and generalized
least-squares regression variants over GLM and GAM methods to MARS and kernel
regression routines. We justify the combinability of their regression
ingredients in a theoretical discourse. Further, we illustrate the approaches
in slightly disguised real-world experiments and perform comprehensive
out-of-sample tests.
| stat |
High Dimensional M-Estimation with Missing Outcomes: A Semi-Parametric
Framework | We consider high dimensional $M$-estimation in settings where the response
$Y$ is possibly missing at random and the covariates $\mathbf{X} \in
\mathbb{R}^p$ can be high dimensional compared to the sample size $n$. The
parameter of interest $\boldsymbol{\theta}_0 \in \mathbb{R}^d$ is defined as
the minimizer of the risk of a convex loss, under a fully non-parametric model,
and $\boldsymbol{\theta}_0$ itself is high dimensional which is a key
distinction from existing works. Standard high dimensional regression and
series estimation with possibly misspecified models and missing $Y$ are
included as special cases, as well as their counterparts in causal inference
using 'potential outcomes'.
Assuming $\boldsymbol{\theta}_0$ is $s$-sparse ($s \ll n$), we propose an
$L_1$-regularized debiased and doubly robust (DDR) estimator of
$\boldsymbol{\theta}_0$ based on a high dimensional adaptation of the
traditional double robust (DR) estimator's construction. Under mild tail
assumptions and arbitrarily chosen (working) models for the propensity score
(PS) and the outcome regression (OR) estimators, satisfying only some
high-level conditions, we establish finite sample performance bounds for the
DDR estimator showing its (optimal) $L_2$ error rate to be $\sqrt{s (\log d)/
n}$ when both models are correct, and its consistency and DR properties when
only one of them is correct. Further, when both the models are correct, we
propose a desparsified version of our DDR estimator that satisfies an
asymptotic linear expansion and facilitates inference on low dimensional
components of $\boldsymbol{\theta}_0$. Finally, we discuss various of choices
of high dimensional parametric/semi-parametric working models for the PS and OR
estimators. All results are validated via detailed simulations.
| stat |
Overview description of the Gambian GABECE Educational Data and
associated algorithms and unsupervized learning process | As the first paper of a series of exploratory analysis and statistical
investigation works on the Gambian \textit{GABECE} data based on a variety of
statistical tools, we wish to begin with a thorough unsupervised learning
process through descriptive and exploratory methods. This will lead to a
variety of discoveries and hypotheses that will direct future research works
related to this data.
| stat |
Efficient and powerful equivalency test on combined mean and variance
with application to diagnostic device comparison studies | In medical device comparison studies, equivalency test is commonly used to
demonstrate two measurement methods agree up to a pre-specified performance
goal based on the paired repeated measures. Such equivalency test often
involves controlling the absolute differences that depend on both the mean and
variance parameters, and poses some challenges for statistical analysis. For
example, for the oximetry comparison study that motivates our research, FDA has
clear guidelines approving an investigational pulse oximeter in comparison to a
standard oximeter via testing the root mean squares (RMS), a composite measure
of both mean and variance parameters. For the hypothesis testing of this
composite measure, existing methods have been either exploratory or relying on
the large-sample normal approximation with conservative and unsatisfactory
performance. We develop a novel generalized pivotal test to rigorously and
accurately test the system equivalency based on RMS. The proposed method has
well-controlled type I error and favorable performance in our extensive
numerical studies. When analyzing data from an oximetry comparison study,
aiming to demonstrate performance equivalency between an FDA-cleared oximetry
system and an investigational system, our proposed method resulted in a highly
significant test result strongly supporting the system equivalency. We also
provide efficient R programs for the proposed method in a publicly available R
package. Considering that many practical equivalency studies of diagnostic
devices are of small to medium sizes, our proposed method and software timely
bridge an existing gap in the field.
| stat |
Distribution free testing for linear regression. Extension to general
parametric regression | Recently a distribution free approach for testing parametric hypotheses based
on unitary transformations has been suggested in \cite{Khm13, Khm16, Khm17} and
further studied in \cite{Ngu17} and \cite{Rob19}. In this note we show that the
transformation takes extremely simple form in distribution free testing of
linear regression. Then we extend it to general parametric regression with
vector-valued covariates.
| stat |
Distributed Variational Bayesian Algorithms Over Sensor Networks | Distributed inference/estimation in Bayesian framework in the context of
sensor networks has recently received much attention due to its broad
applicability. The variational Bayesian (VB) algorithm is a technique for
approximating intractable integrals arising in Bayesian inference. In this
paper, we propose two novel distributed VB algorithms for general Bayesian
inference problem, which can be applied to a very general class of
conjugate-exponential models. In the first approach, the global natural
parameters at each node are optimized using a stochastic natural gradient that
utilizes the Riemannian geometry of the approximation space, followed by an
information diffusion step for cooperation with the neighbors. In the second
method, a constrained optimization formulation for distributed estimation is
established in natural parameter space and solved by alternating direction
method of multipliers (ADMM). An application of the distributed
inference/estimation of a Bayesian Gaussian mixture model is then presented, to
evaluate the effectiveness of the proposed algorithms. Simulations on both
synthetic and real datasets demonstrate that the proposed algorithms have
excellent performance, which are almost as good as the corresponding
centralized VB algorithm relying on all data available in a fusion center.
| stat |
Variational f-divergence Minimization | Probabilistic models are often trained by maximum likelihood, which
corresponds to minimizing a specific f-divergence between the model and data
distribution. In light of recent successes in training Generative Adversarial
Networks, alternative non-likelihood training criteria have been proposed.
Whilst not necessarily statistically efficient, these alternatives may better
match user requirements such as sharp image generation. A general variational
method for training probabilistic latent variable models using maximum
likelihood is well established; however, how to train latent variable models
using other f-divergences is comparatively unknown. We discuss a variational
approach that, when combined with the recently introduced Spread Divergence,
can be applied to train a large class of latent variable models using any
f-divergence.
| stat |
Adversarial Robustness Guarantees for Classification with Gaussian
Processes | We investigate adversarial robustness of Gaussian Process Classification
(GPC) models. Given a compact subset of the input space $T\subseteq
\mathbb{R}^d$ enclosing a test point $x^*$ and a GPC trained on a dataset
$\mathcal{D}$, we aim to compute the minimum and the maximum classification
probability for the GPC over all the points in $T$. In order to do so, we show
how functions lower- and upper-bounding the GPC output in $T$ can be derived,
and implement those in a branch and bound optimisation algorithm. For any error
threshold $\epsilon > 0$ selected a priori, we show that our algorithm is
guaranteed to reach values $\epsilon$-close to the actual values in finitely
many iterations. We apply our method to investigate the robustness of GPC
models on a 2D synthetic dataset, the SPAM dataset and a subset of the MNIST
dataset, providing comparisons of different GPC training techniques, and show
how our method can be used for interpretability analysis. Our empirical
analysis suggests that GPC robustness increases with more accurate posterior
estimation.
| stat |
Estimating and comparing adverse event probabilities in the presence of
varying follow-up times and competing events | Safety analyses in terms of adverse events (AEs) are an important aspect of
benefit-risk assessments of therapies. Compared to efficacy analyses AE
analyses are often rather simplistic. The probability of an AE of a specific
type is typically estimated by the incidence proportion, sometimes the
incidence density or the Kaplan-Meier estimator are proposed. But these
analyses either do not account for censoring, rely on a too restrictive
parametric model, or ignore competing events. With the non-parametric
Aalen-Johansen estimator as the gold-standard, these potential sources of bias
are investigated in a data example from oncology and in simulations, both in
the one-sample and in the two-sample case. As the estimators may have large
variances at the end of follow-up, the estimators are not only compared at the
maximal event time but also at two quantiles of the observed times. To date,
consequences for safety comparisons have hardly been investigated in the
literature. The impact of using different estimators for group comparisons is
unclear, as, for example, the ratio of two both underestimating or
overestimating estimators may or may not be comparable to the ratio of the
gold-standard estimator. Therefore, the ratio of the AE probabilities is also
calculated based on different approaches. By simulations investigating constant
and non-constant hazards, different censoring mechanisms and event frequencies,
we show that ignoring competing events is more of a problem than falsely
assuming constant hazards by use of the incidence density and that the choice
of the AE probability estimator is crucial for group comparisons.
| stat |
Graph-Based Two-Sample Tests for Data with Repeated Observations | In the regime of two-sample comparison, tests based on a graph constructed on
observations by utilizing similarity information among them is gaining
attention due to their flexibility and good performances for
high-dimensional/non-Euclidean data. However, when there are repeated
observations, these graph-based tests could be problematic as they are
versatile to the choice of the similarity graph. We propose extended
graph-based test statistics to resolve this problem. The analytic p-value
approximations to these extended graph-based tests are derived to facilitate
the application of these tests to large datasets. The new tests are illustrated
in the analysis of a phone-call network dataset. All tests are implemented in
an R package gTests.
| stat |
Multilevel regression with poststratification for the national level
Viber/Street poll on the 2020 presidential election in Belarus | Independent sociological polls are forbidden in Belarus. Online polls
performed without sound scientific rigour do not yield representative results.
Yet, both inside and outside Belarus it is of great importance to obtain
precise estimates of the ratings of all candidates. These ratings could
function as reliable proxies for the election's outcomes. We conduct an
independent poll based on the combination of the data collected via Viber and
on the streets of Belarus. The Viber and the street data samples consist of
almost 45000 and 1150 unique observations respectively. Bayesian regressions
with poststratification were build to estimate ratings of the candidates and
rates of early voting turnout for the population as a whole and within various
focus subgroups. We show that both the officially announced results of the
election and early voting rates are highly improbable. With a probability of at
least 95%, Sviatlana Tikhanouskaya's rating lies between 75% and 80%, whereas
Aliaksandr Lukashenka's rating lies between 13% and 18% and early voting rate
predicted by the method ranges from 9% to 13% of those who took part in the
election. These results contradict the officially announced outcomes, which are
10.12%, 80.11%, and 49.54% respectively and lie far outside even the 99.9%
credible intervals predicted by our model. The only marginal groups of people
where the upper bounds of the 99.9% credible intervals of the rating of
Lukashenka are above 50% are people older than 60 and uneducated people. For
all other marginal subgroups, including rural residents, even the upper bounds
of 99.9% credible intervals for Lukashenka are far below 50%. The same is true
for the population as a whole. Thus, with a probability of at least 99.9%
Lukashenka could not have had enough electoral support to win the 2020
presidential election in Belarus.
| stat |
Predicting colorectal polyp recurrence using time-to-event analysis of
medical records | Identifying patient characteristics that influence the rate of colorectal
polyp recurrence can provide important insights into which patients are at
higher risk for recurrence. We used natural language processing to extract
polyp morphological characteristics from 953 polyp-presenting patients'
electronic medical records. We used subsequent colonoscopy reports to examine
how the time to polyp recurrence (731 patients experienced recurrence) is
influenced by these characteristics as well as anthropometric features using
Kaplan-Meier curves, Cox proportional hazards modeling, and random survival
forest models. We found that the rate of recurrence differed significantly by
polyp size, number, and location and patient smoking status. Additionally,
right-sided colon polyps increased recurrence risk by 30% compared to
left-sided polyps. History of tobacco use increased polyp recurrence risk by
20% compared to never-users. A random survival forest model showed an AUC of
0.65 and identified several other predictive variables, which can inform
development of personalized polyp surveillance plans.
| stat |
e4clim 1.0 : The Energy for CLimate Integrated Model: Description and
Application to Italy | We develop an open-source Python software integrating flexibility needs from
Variable Renewable Energies (VREs) in the development of regional energy mixes.
It provides a flexible and extensible tool to researchers/engineers, and for
education/outreach. It aims at evaluating and optimizing energy deployment
strategies with high shares of VRE; assessing the impact of new technologies
and of climate variability; conducting sensitivity studies. Specifically, to
limit the algorithm's complexity, we avoid solving a full-mix cost-minimization
problem by taking the mean and variance of the renewable production-demand
ratio as proxies to balance services. Second, observations of VRE technologies
being typically too short or nonexistent, the hourly demand and production are
estimated from climate time-series and fitted to available observations. We
illustrate e4clim's potential with an optimal recommissioning-study of the 2015
Italian PV-wind mix testing different climate-data sources and strategies and
assessing the impact of climate variability and the robustness of the results.
| stat |
Bayesian statistical analysis of hydrogeochemical data using point
processes: a new tool for source detection in multicomponent fluid mixtures | Hydrogeochemical data may be seen as a point cloud in a multi-dimensional
space. Each dimension of this space represents a hydrogeochemical parameter
(i.e. salinity, solute concentration, concentration ratio, isotopic
composition...). While the composition of many geological fluids is controlled
by mixing between multiple sources, a key question related to hydrogeochemical
data set is the detection of the sources. By looking at the hydrogeochemical
data as spatial data, this paper presents a new solution to the source
detection problem that is based on point processes. Results are shown on
simulated and real data from geothermal fluids.
| stat |
Variational approach for learning Markov processes from time series data | Inference, prediction and control of complex dynamical systems from time
series is important in many areas, including financial markets, power grid
management, climate and weather modeling, or molecular dynamics. The analysis
of such highly nonlinear dynamical systems is facilitated by the fact that we
can often find a (generally nonlinear) transformation of the system coordinates
to features in which the dynamics can be excellently approximated by a linear
Markovian model. Moreover, the large number of system variables often change
collectively on large time- and length-scales, facilitating a low-dimensional
analysis in feature space. In this paper, we introduce a variational approach
for Markov processes (VAMP) that allows us to find optimal feature mappings and
optimal Markovian models of the dynamics from given time series data. The key
insight is that the best linear model can be obtained from the top singular
components of the Koopman operator. This leads to the definition of a family of
score functions called VAMP-r which can be calculated from data, and can be
employed to optimize a Markovian model. In addition, based on the relationship
between the variational scores and approximation errors of Koopman operators,
we propose a new VAMP-E score, which can be applied to cross-validation for
hyper-parameter optimization and model selection in VAMP. VAMP is valid for
both reversible and nonreversible processes and for stationary and
non-stationary processes or realizations.
| stat |
Model-Based Clustering of Nonparametric Weighted Networks with
Application to Water Pollution Analysis | Water pollution is a major global environmental problem, and it poses a great
environmental risk to public health and biological diversity. This work is
motivated by assessing the potential environmental threat of coal mining
through increased sulfate concentrations in river networks, which do not belong
to any simple parametric distribution. However, existing network models mainly
focus on binary or discrete networks and weighted networks with known
parametric weight distributions. We propose a principled nonparametric weighted
network model based on exponential-family random graph models and local
likelihood estimation and study its model-based clustering with application to
large-scale water pollution network analysis. We do not require any parametric
distribution assumption on network weights. The proposed method greatly extends
the methodology and applicability of statistical network models. Furthermore,
it is scalable to large and complex networks in large-scale environmental
studies. The power of our proposed methods is demonstrated in simulation
studies and a real application to sulfate pollution network analysis in Ohio
watershed located in Pennsylvania, United States.
| stat |
GraphNVP: An Invertible Flow Model for Generating Molecular Graphs | We propose GraphNVP, the first invertible, normalizing flow-based molecular
graph generation model. We decompose the generation of a graph into two steps:
generation of (i) an adjacency tensor and (ii) node attributes. This
decomposition yields the exact likelihood maximization on graph-structured
data, combined with two novel reversible flows. We empirically demonstrate that
our model efficiently generates valid molecular graphs with almost no
duplicated molecules. In addition, we observe that the learned latent space can
be used to generate molecules with desired chemical properties.
| stat |
Diffusion Approximations for a Class of Sequential Testing Problems | We consider a decision maker who must choose an action in order to maximize a
reward function that depends also on an unknown parameter {\Theta}. The
decision maker can delay taking the action in order to experiment and gather
additional information on {\Theta}. We model the decision maker's problem using
a Bayesian sequential experimentation framework and use dynamic programming and
diffusion-asymptotic analysis to solve it. For that, we scale our problem in a
way that both the average number of experiments that is conducted per unit of
time is large and the informativeness of each individual experiment is low.
Under such regime, we derive a diffusion approximation for the sequential
experimentation problem, which provides a number of important insights about
the nature of the problem and its solution. Our solution method also shows that
the complexity of the problem grows only quadratically with the cardinality of
the set of actions from which the decision maker can choose. We illustrate our
methodology and results using a concrete application in the context of
assortment selection and new product introduction. Specifically, we study the
problem of a seller who wants to select an optimal assortment of products to
launch into the marketplace and is uncertain about consumers' preferences.
Motivated by emerging practices in e-commerce, we assume that the seller is
able to use a crowdvoting system to learn these preferences before a final
assortment decision is made. In this context, we undertake an extensive
numerical analysis to assess the value of learning and demonstrate the
effectiveness and robustness of the heuristics derived from the diffusion
approximation.
| stat |
Simultaneous Dimensionality and Complexity Model Selection for Spectral
Graph Clustering | Our problem of interest is to cluster vertices of a graph by identifying
underlying community structure. Among various vertex clustering approaches,
spectral clustering is one of the most popular methods because it is easy to
implement while often outperforming more traditional clustering algorithms.
However, there are two inherent model selection problems in spectral
clustering, namely estimating both the embedding dimension and number of
clusters. This paper attempts to address the issue by establishing a novel
model selection framework specifically for vertex clustering on graphs under a
stochastic block model. The first contribution is a probabilistic model which
approximates the distribution of the extended spectral embedding of a graph.
The model is constructed based on a theoretical result of asymptotic normality
of the informative part of the embedding, and on a simulation result providing
a conjecture for the limiting behavior of the redundant part of the embedding.
The second contribution is a simultaneous model selection framework. In
contrast with the traditional approaches, our model selection procedure
estimates embedding dimension and number of clusters simultaneously. Based on
our conjectured distributional model, a theorem on the consistency of the
estimates of model parameters is presented, providing support for the validity
of our method. Algorithms for our simultaneous model selection for vertex
clustering are proposed, demonstrating superior performance in simulation
experiments. We illustrate our method via application to a collection of brain
graphs.
| stat |
Minimum Description Length Revisited | This is an up-to-date introduction to and overview of the Minimum Description
Length (MDL) Principle, a theory of inductive inference that can be applied to
general problems in statistics, machine learning and pattern recognition. While
MDL was originally based on data compression ideas, this introduction can be
read without any knowledge thereof. It takes into account all major
developments since 2007, the last time an extensive overview was written. These
include new methods for model selection and averaging and hypothesis testing,
as well as the first completely general definition of {\em MDL estimators}.
Incorporating these developments, MDL can be seen as a powerful extension of
both penalized likelihood and Bayesian approaches, in which penalization
functions and prior distributions are replaced by more general luckiness
functions, average-case methodology is replaced by a more robust worst-case
approach, and in which methods classically viewed as highly distinct, such as
AIC vs BIC and cross-validation vs Bayes can, to a large extent, be viewed from
a unified perspective.
| stat |
On the Monotonicity of a Nondifferentially Mismeasured Binary Confounder | Suppose that we are interested in the average causal effect of a binary
treatment on an outcome when this relationship is confounded by a binary
confounder. Suppose that the confounder is unobserved but a nondifferential
proxy of it is observed. We show that, under certain monotonicity assumption
that is empirically verifiable, adjusting for the proxy produces a measure of
the effect that is between the unadjusted and the true measures.
| stat |
A Two-Sample Robust Bayesian Mendelian Randomization Method Accounting
for Linkage Disequilibrium and Idiosyncratic Pleiotropy with Applications to
the COVID-19 Outcome | Mendelian randomization (MR) is a statistical method exploiting genetic
variants as instrumental variables to estimate the causal effect of modifiable
risk factors on an outcome of interest. Despite wide uses of various popular
two-sample MR methods based on genome-wide association study summary level
data, however, those methods could suffer from potential power loss or/and
biased inference when the chosen genetic variants are in linkage disequilibrium
(LD), and also have relatively large direct effects on the outcome whose
distribution might be heavy-tailed which is commonly referred to as the
idiosyncratic pleiotropy phenomenon. To resolve those two issues, we propose a
novel Robust Bayesian Mendelian Randomization (RBMR) model that uses the more
robust multivariate generalized t-distribution to model such direct effects in
a probabilistic model framework which can also incorporate the LD structure
explicitly. The generalized t-distribution can be represented as a Gaussian
scaled mixture so that our model parameters can be estimated by the EM-type
algorithms. We compute the standard errors by calibrating the evidence lower
bound using the likelihood ratio test. Through extensive simulation studies, we
show that our RBMR has robust performance compared to other competing methods.
We also apply our RBMR method to two benchmark data sets and find that RBMR has
smaller bias and standard errors. Using our proposed RBMR method, we find that
coronary artery disease is associated with increased risk of critically ill
coronavirus disease 2019 (COVID-19). We also develop a user-friendly R package
RBMR for public use.
| stat |
PROFIT: Projection-based Test in Longitudinal Functional Data | In many modern applications, a dependent functional response is observed for
each subject over repeated time, leading to longitudinal functional data. In
this paper, we propose a novel statistical procedure to test whether the mean
function varies over time. Our approach relies on reducing the dimension of the
response using data-driven orthogonal projections and it employs a
likelihood-based hypothesis testing. We investigate the methodology
theoretically and discuss a computationally efficient implementation. The
proposed test maintains the type I error rate, and shows excellent power to
detect departures from the null hypothesis in finite sample simulation studies.
We apply our method to the longitudinal diffusion tensor imaging study of
multiple sclerosis (MS) patients to formally assess whether the brain's health
tissue, as summarized by fractional anisotropy (FA) profile, degrades over time
during the study period.
| stat |
Predicting Mortality Risk in Viral and Unspecified Pneumonia to Assist
Clinicians with COVID-19 ECMO Planning | Respiratory complications due to coronavirus disease COVID-19 have claimed
tens of thousands of lives in 2020. Many cases of COVID-19 escalate from Severe
Acute Respiratory Syndrome (SARS-CoV-2) to viral pneumonia to acute respiratory
distress syndrome (ARDS) to death. Extracorporeal membranous oxygenation (ECMO)
is a life-sustaining oxygenation and ventilation therapy that may be used for
patients with severe ARDS when mechanical ventilation is insufficient to
sustain life. While early planning and surgical cannulation for ECMO can
increase survival, clinicians report the lack of a risk score hinders these
efforts. In this work, we leverage machine learning techniques to develop the
PEER score, used to highlight critically ill patients with viral or unspecified
pneumonia at high risk of mortality or decompensation in a subpopulation
eligible for ECMO. The PEER score is validated on two large, publicly available
critical care databases and predicts mortality at least as well as other
existing risk scores. Stratifying our cohorts into low-risk and high-risk
groups, we find that the high-risk group also has a higher proportion of
decompensation indicators such as vasopressor and ventilator use. Finally, the
PEER score is provided in the form of a nomogram for direct calculation of
patient risk, and can be used to highlight at-risk patients among critical care
patients eligible for ECMO.
| stat |
Detected changes in precipitation extremes at their native scales
derived from in situ measurements | The gridding of daily accumulated precipitation -- especially extremes --
from ground-based station observations is problematic due to the fractal nature
of precipitation, and therefore estimates of long period return values and
their changes based on such gridded daily data sets are generally
underestimated. In this paper, we characterize high-resolution changes in
observed extreme precipitation from 1950 to 2017 for the contiguous United
States (CONUS) based on in situ measurements only. Our analysis utilizes
spatial statistical methods that allow us to derive gridded estimates that do
not smooth extreme daily measurements and are consistent with statistics from
the original station data while increasing the resulting signal to noise ratio.
Furthermore, we use a robust statistical technique to identify significant
pointwise changes in the climatology of extreme precipitation while carefully
controlling the rate of false positives. We present and discuss seasonal
changes in the statistics of extreme precipitation: the largest and most
spatially-coherent pointwise changes are in fall (SON), with approximately 33%
of CONUS exhibiting significant changes (in an absolute sense). Other seasons
display very few meaningful pointwise changes (in either a relative or absolute
sense), illustrating the difficulty in detecting pointwise changes in extreme
precipitation based on in situ measurements. While our main result involves
seasonal changes, we also present and discuss annual changes in the statistics
of extreme precipitation. In this paper we only seek to detect changes over
time and leave attribution of the underlying causes of these changes for future
work.
| stat |
Gaussian processes with linear operator inequality constraints | This paper presents an approach for constrained Gaussian Process (GP)
regression where we assume that a set of linear transformations of the process
are bounded. It is motivated by machine learning applications for
high-consequence engineering systems, where this kind of information is often
made available from phenomenological knowledge. We consider a GP $f$ over
functions on $\mathcal{X} \subset \mathbb{R}^{n}$ taking values in
$\mathbb{R}$, where the process $\mathcal{L}f$ is still Gaussian when
$\mathcal{L}$ is a linear operator. Our goal is to model $f$ under the
constraint that realizations of $\mathcal{L}f$ are confined to a convex set of
functions. In particular, we require that $a \leq \mathcal{L}f \leq b$, given
two functions $a$ and $b$ where $a < b$ pointwise. This formulation provides a
consistent way of encoding multiple linear constraints, such as
shape-constraints based on e.g. boundedness, monotonicity or convexity. We
adopt the approach of using a sufficiently dense set of virtual observation
locations where the constraint is required to hold, and derive the exact
posterior for a conjugate likelihood. The results needed for stable numerical
implementation are derived, together with an efficient sampling scheme for
estimating the posterior process.
| stat |
A finite sample analysis of the benign overfitting phenomenon for ridge
function estimation | Recent extensive numerical experiments in high scale machine learning have
allowed to uncover a quite counterintuitive phase transition, as a function of
the ratio between the sample size and the number of parameters in the model. As
the number of parameters $p$ approaches the sample size $n$, the generalisation
error (a.k.a. testing error) increases, but in many cases, it starts decreasing
again past the threshold $p=n$. This surprising phenomenon, brought to the
theoretical community attention in \cite{belkin2019reconciling}, has been
thoroughly investigated lately, more specifically for simpler models than deep
neural networks, such as the linear model when the parameter is taken to be the
minimum norm solution to the least-square problem, mostly in the asymptotic
regime when $p$ and $n$ tend to $+\infty$; see e.g. \cite{hastie2019surprises}.
In the present paper, we propose a finite sample analysis of non-linear models
of \textit{ridge} type, where we investigate the \textit{overparametrised
regime} of the double descent phenomenon for both the \textit{estimation
problem} and the \textit{prediction} problem. Our results provide a precise
analysis of the distance of the best estimator from the true parameter as well
as a generalisation bound which complements recent works of
\cite{bartlett2020benign} and \cite{chinot2020benign}. Our analysis is based on
efficient but elementary tools closely related to the continuous Newton method
\cite{neuberger2007continuous}.
| stat |
Data Preprocessing to Mitigate Bias with Boosted Fair Mollifiers | In a recent paper, Celis et al. (2020) introduced a new approach to fairness
that corrects the data distribution itself. The approach is computationally
appealing, but its approximation guarantees with respect to the target
distribution can be quite loose as they need to rely on a (typically limited)
number of constraints on data-based aggregated statistics; also resulting in a
fairness guarantee which can be data dependent.
Our paper makes use of a mathematical object recently introduced in privacy
-- mollifiers of distributions -- and a popular approach to machine learning --
boosting -- to get an approach in the same lineage as Celis et al. but without
the same impediments, including in particular, better guarantees in terms of
accuracy and finer guarantees in terms of fairness. The approach involves
learning the sufficient statistics of an exponential family. When the training
data is tabular, the sufficient statistics can be defined by decision trees
whose interpretability can provide clues on the source of (un)fairness.
Experiments display the quality of the results for simulated and real-world
data.
| stat |
Parameter Estimation of Heavy-Tailed AR Model with Missing Data via
Stochastic EM | The autoregressive (AR) model is a widely used model to understand time
series data. Traditionally, the innovation noise of the AR is modeled as
Gaussian. However, many time series applications, for example, financial time
series data, are non-Gaussian, therefore, the AR model with more general
heavy-tailed innovations is preferred. Another issue that frequently occurs in
time series is missing values, due to system data record failure or unexpected
data loss. Although there are numerous works about Gaussian AR time series with
missing values, as far as we know, there does not exist any work addressing the
issue of missing data for the heavy-tailed AR model. In this paper, we consider
this issue for the first time, and propose an efficient framework for parameter
estimation from incomplete heavy-tailed time series based on a stochastic
approximation expectation maximization (SAEM) coupled with a Markov Chain Monte
Carlo (MCMC) procedure. The proposed algorithm is computationally cheap and
easy to implement. The convergence of the proposed algorithm to a stationary
point of the observed data likelihood is rigorously proved. Extensive
simulations and real datasets analyses demonstrate the efficacy of the proposed
framework.
| stat |
Principal components variable importance reconstruction (PC-VIR):
Exploring predictive importance in multicollinear acoustic speech data | This paper presents a method of exploring the relative predictive importance
of individual variables in multicollinear data sets at three levels of
significance: strong importance, moderate importance, and no importance.
Implementation of Bonferroni adjustment to control for Type I error in the
method is described, and results with and without the correction are compared.
An example of the method in binary logistic modeling is demonstrated by using a
set of 20 acoustic features to discriminate vocalic nasality in the speech of
six speakers of the Mixean variety of Low Navarrese Basque. Validation of the
method is presented by comparing the direction of significant effects to those
observed in separate logistic mixed effects models, as well as goodness of fit
and prediction accuracy compared to partial least squares logistic regression.
The results show that the proposed method yields: (1) similar, but more
conservative estimates in comparison to separate logistic regression models,
(2) models that fit data as well as partial least squares methods, and (3)
predictions for new data that are as accurate as partial least squares methods.
| stat |
High-dimensional Model-assisted Inference for Local Average Treatment
Effects with Instrumental Variables | Consider the problem of estimating the local average treatment effect with an
instrument variable, where the instrument unconfoundedness holds after
adjusting for a set of measured covariates. Several unknown functions of the
covariates need to be estimated through regression models, such as instrument
propensity score and treatment and outcome regression models. We develop a
computationally tractable method in high-dimensional settings where the numbers
of regression terms are close to or larger than the sample size. Our method
exploits regularized calibrated estimation, which involves Lasso penalties but
carefully chosen loss functions for estimating coefficient vectors in these
regression models, and then employs a doubly robust estimator for the treatment
parameter through augmented inverse probability weighting. We provide rigorous
theoretical analysis to show that the resulting Wald confidence intervals are
valid for the treatment parameter under suitable sparsity conditions if the
instrument propensity score model is correctly specified, but the treatment and
outcome regression models may be misspecified. For existing high-dimensional
methods, valid confidence intervals are obtained for the treatment parameter if
all three models are correctly specified. We evaluate the proposed methods via
extensive simulation studies and an empirical application to estimate the
returns to education.
| stat |
Online Learning Algorithms for Quaternion ARMA Model | In this paper, we address the problem of adaptive learning for autoregressive
moving average (ARMA) model in the quaternion domain. By transforming the
original learning problem into a full information optimization task without
explicit noise terms, and then solving the optimization problem using the
gradient descent and the Newton analogues, we obtain two online learning
algorithms for the quaternion ARMA. Furthermore, regret bound analysis
accounting for the specific properties of quaternion algebra is presented,
which proves that the performance of the online algorithms asymptotically
approaches that of the best quaternion ARMA model in hindsight.
| stat |
A Unified Framework for Random Forest Prediction Error Estimation | We introduce a unified framework for random forest prediction error
estimation based on a novel estimator of the conditional prediction error
distribution function. Our framework enables simple plug-in estimation of key
prediction uncertainty metrics, including conditional mean squared prediction
errors, conditional biases, and conditional quantiles, for random forests and
many variants. Our approach is especially well-adapted for prediction interval
estimation; we show via simulations that our proposed prediction intervals are
competitive with, and in some settings outperform, existing methods. To
establish theoretical grounding for our framework, we prove pointwise uniform
consistency of a more stringent version of our estimator of the conditional
prediction error distribution function. The estimators introduced here are
implemented in the R package forestError.
| stat |
Conditional Variational Inference with Adaptive Truncation for Bayesian
Nonparametric Models | The scalable inference for Bayesian nonparametric models with big data is
still challenging. Current variational inference methods fail to characterise
the correlation structure among latent variables due to the mean-field setting
and cannot infer the true posterior dimension because of the universal
truncation. To overcome these limitations, we build a general framework to
infer Bayesian nonparametric models by maximising the proposed nonparametric
evidence lower bound, and then develop a novel approach by combining Monte
Carlo sampling and stochastic variational inference framework. Our method has
several advantages over the traditional online variational inference method.
First, it achieves a smaller divergence between variational distributions and
the true posterior by factorising variational distributions under the
conditional setting instead of the mean-field setting to capture the
correlation pattern. Second, it reduces the risk of underfitting or overfitting
by truncating the dimension adaptively rather than using a prespecified
truncated dimension for all latent variables. Third, it reduces the
computational complexity by approximating the posterior functionally instead of
updating the stick-breaking parameters individually. We apply the proposed
method on hierarchical Dirichlet process and gamma--Dirichlet process models,
two essential Bayesian nonparametric models in topic analysis. The empirical
study on three large datasets including arXiv, New York Times and Wikipedia
reveals that our proposed method substantially outperforms its competitor in
terms of lower perplexity and much clearer topic-words clustering.
| stat |
Adaptive MCMC for Generalized Method of Moments with Many Moment
Conditions | A generalized method of moments (GMM) estimator is unreliable for a large
number of moment conditions, that is, it is comparable, or larger than the
sample size. While classical GMM literature proposes several provisions to this
problem, its Bayesian counterpart (i.e., Bayesian inference using a GMM
criterion as a quasi-likelihood) almost totally ignores it. This study bridges
this gap by proposing an adaptive Markov Chain Monte Carlo (MCMC) approach to a
GMM inference with many moment conditions. Particularly, this study focuses on
the adaptive tuning of a weighting matrix on the fly. Our proposal consists of
two elements. The first is the use of the nonparametric eigenvalue-regularized
precision matrix estimator, which contributes to numerical stability. The
second is the random update of a weighting matrix, which substantially reduces
computational cost, while maintaining the accuracy of the estimation. We then
present a simulation study and real data application to compare the performance
of the proposed approach with existing approaches.
| stat |
Multilevel Bootstrap Particle Filter | We consider situations where the applicability of sequential Monte Carlo
particle filters is compromised due to the expensive evaluation of the particle
weights. To alleviate this problem, we propose a new particle filter algorithm
based on the multilevel approach. We show that the resulting multilevel
bootstrap particle filter (MLBPF) retains the strong law of large numbers as
well as the central limit theorem of classical particle filters under mild
conditions. Our numerical experiments demonstrate up to 85\% reduction in
computation time compared to the classical bootstrap particle filter, in
certain settings. While it should be acknowledged that this reduction is highly
application dependent, and a similar gain should not be expected for all
applications across the board, we believe that this substantial improvement in
certain settings makes MLBPF an important addition to the family of sequential
Monte Carlo methods.
| stat |
On Last-Layer Algorithms for Classification: Decoupling Representation
from Uncertainty Estimation | Uncertainty quantification for deep learning is a challenging open problem.
Bayesian statistics offer a mathematically grounded framework to reason about
uncertainties; however, approximate posteriors for modern neural networks still
require prohibitive computational costs. We propose a family of algorithms
which split the classification task into two stages: representation learning
and uncertainty estimation. We compare four specific instances, where
uncertainty estimation is performed via either an ensemble of Stochastic
Gradient Descent or Stochastic Gradient Langevin Dynamics snapshots, an
ensemble of bootstrapped logistic regressions, or via a number of Monte Carlo
Dropout passes. We evaluate their performance in terms of \emph{selective}
classification (risk-coverage), and their ability to detect out-of-distribution
samples. Our experiments suggest there is limited value in adding multiple
uncertainty layers to deep classifiers, and we observe that these simple
methods strongly outperform a vanilla point-estimate SGD in some complex
benchmarks like ImageNet.
| stat |
The scale transformed power prior for use with historical data from a
different outcome model | We develop the scale transformed power prior for settings where historical
and current data involve different data types, such as binary and continuous
data, respectively. This situation arises often in clinical trials, for
example, when historical data involve binary responses and the current data
involve time-to-event or some other type of continuous or discrete outcome. The
power prior proposed by Ibrahim and Chen (2000) does not address the issue of
different data types. Herein, we develop a new type of power prior, which we
call the scale transformed power prior (straPP). The straPP is constructed by
transforming the power prior for the historical data by rescaling the parameter
using a function of the Fisher information matrices for the historical and
current data models, thereby shifting the scale of the parameter vector from
that of the historical to that of the current data. Examples are presented to
motivate the need for a scale transformation and simulation studies are
presented to illustrate the performance advantages of the straPP over the power
prior and other informative and non-informative priors. A real dataset from a
clinical trial undertaken to study a novel transitional care model for stroke
survivors is used to illustrate the methodology.
| stat |
Global Capacity Measures for Deep ReLU Networks via Path Sampling | Classical results on the statistical complexity of linear models have
commonly identified the norm of the weights $\|w\|$ as a fundamental capacity
measure. Generalizations of this measure to the setting of deep networks have
been varied, though a frequently identified quantity is the product of weight
norms of each layer. In this work, we show that for a large class of networks
possessing a positive homogeneity property, similar bounds may be obtained
instead in terms of the norm of the product of weights. Our proof technique
generalizes a recently proposed sampling argument, which allows us to
demonstrate the existence of sparse approximants of positive homogeneous
networks. This yields covering number bounds, which can be converted to
generalization bounds for multi-class classification that are comparable to,
and in certain cases improve upon, existing results in the literature. Finally,
we investigate our sampling procedure empirically, which yields results
consistent with our theory.
| stat |
A multi-state model incorporating estimation of excess hazards and
multiple time scales | As cancer patient survival improves, late effects from treatment are becoming
the next clinical challenge. Chemotherapy and radiotherapy, for example,
potentially increase the risk of both morbidity and mortality from second
malignancies and cardiovascular disease. To provide clinically relevant
population-level measures of late effects, it is of importance to (1)
simultaneously estimate the risks of both morbidity and mortality, (2)
partition these risks into the component expected in the absence of cancer and
the component due to the cancer and its treatment, and (3) incorporate the
multiple time scales of attained age, calendar time, and time since diagnosis.
Multi-state models provide a framework for simultaneously studying morbidity
and mortality, but do not solve the problem of partitioning the risks. However,
this partitioning can be achieved by applying a relative survival framework, by
allowing is to directly quantify the excess risk. This paper proposes a
combination of these two frameworks, providing one approach to address (1)-(3).
Using recently developed methods in multi-state modeling, we incorporate
estimation of excess hazards into a multi-state model. Both intermediate and
absorbing state risks can be partitioned and different transitions are allowed
to have different and/or multiple time scales. We illustrate our approach using
data on Hodgkin lymphoma patients and excess risk of diseases of the
circulatory system, and provide user-friendly Stata software with accompanying
example code.
| stat |
A Two-Stage Approach to Multivariate Linear Regression with Sparsely
Mismatched Data | A tacit assumption in linear regression is that (response, predictor)-pairs
correspond to identical observational units. A series of recent works have
studied scenarios in which this assumption is violated under terms such as
``Unlabeled Sensing and ``Regression with Unknown Permutation''. In this paper,
we study the setup of multiple response variables and a notion of mismatches
that generalizes permutations in order to allow for missing matches as well as
for one-to-many matches. A two-stage method is proposed under the assumption
that most pairs are correctly matched. In the first stage, the regression
parameter is estimated by handling mismatches as contaminations, and
subsequently the generalized permutation is estimated by a basic variant of
matching. The approach is both computationally convenient and equipped with
favorable statistical guarantees. Specifically, it is shown that the conditions
for permutation recovery become considerably less stringent as the number of
responses $m$ per observation increase. Particularly, for $m = \Omega(\log n)$,
the required signal-to-noise ratio no longer depends on the sample size $n$.
Numerical results on synthetic and real data are presented to support the main
findings of our analysis.
| stat |
Latent space models for multiplex networks with shared structure | Latent space models are frequently used for modeling single-layer networks
and include many popular special cases, such as the stochastic block model and
the random dot product graph. However, they are not well-developed for more
complex network structures, which are becoming increasingly common in practice.
Here we propose a new latent space model for multiplex networks: multiple,
heterogeneous networks observed on a shared node set. Multiplex networks can
represent a network sample with shared node labels, a network evolving over
time, or a network with multiple types of edges. The key feature of our model
is that it learns from data how much of the network structure is shared between
layers and pools information across layers as appropriate. We establish
identifiability, develop a fitting procedure using convex optimization in
combination with a nuclear norm penalty, and prove a guarantee of recovery for
the latent positions as long as there is sufficient separation between the
shared and the individual latent subspaces. We compare the model to competing
methods in the literature on simulated networks and on a multiplex network
describing the worldwide trade of agricultural products.
| stat |
Seasonal association between viral causes of hospitalised acute lower
respiratory infections and meteorological factors in China: a retrospective
study | Acute lower respiratory infections caused by respiratory viruses are common
and persistent infectious diseases worldwide and in China, which have
pronounced seasonal patterns. Meteorological factors have important roles in
the seasonality of some major viruses. Our aim was to identify the dominant
meteorological factors and to model their effects on common respiratory viruses
in different regions of China. We analysed monthly virus data on patients from
81 sentinel hospitals in 22 provinces in mainland China from 2009 to 2013. The
geographical detector method was used to quantify the explanatory power of each
meteorological factor, individually and interacting in pairs. 28369
hospitalised patients with ALRI were tested, 10387 were positive for at least
one virus, including RSV, influenza virus, PIV, ADV, hBoV, hCoV and hMPV. RSV
and influenza virus had annual peaks in the north and biannual peaks in the
south. PIV and hBoV had higher positive rates in the spring summer months. hMPV
had an annual peak in winter spring, especially in the north. ADV and hCoV
exhibited no clear annual seasonality. Temperature, atmospheric pressure,
vapour pressure, and rainfall had most explanatory power on most respiratory
viruses in each region. Relative humidity was only dominant in the north, but
had no significant explanatory power for most viruses in the south. Hours of
sunlight had significant explanatory power for RSV and influenza virus in the
north, and for most viruses in the south. Wind speed was the only factor with
significant explanatory power for human coronavirus in the south. For all
viruses, interactions between any two of the paired factors resulted in
enhanced explanatory power, either bivariately or non-linearly.
| stat |
Modeling Count Data via Copulas | Copula models have been widely used to model the dependence between
continuous random variables, but modeling count data via copulas has recently
become popular in the statistics literature. Spearman's rho is an appropriate
and effective tool to measure the degree of dependence between two random
variables. In this paper, we derived the population version of Spearman's rho
correlation via copulas when both random variables are discrete. The
closed-form expressions of the Spearman correlation are obtained for some
copulas of simple structure such as Archimedean copulas with different marginal
distributions. We derive the upper bound and the lower bound of the Spearman's
rho for Bernoulli random variables. Then, the proposed Spearman's rho
correlations are compared with their corresponding Kendall's tau values. We
characterize the functional relationship between these two measures of
dependence in some special cases. An extensive simulation study is conducted to
demonstrate the validity of our theoretical results. Finally, we propose a
bivariate copula regression model to analyze the count data of a \emph{cervical
cancer} dataset.
| stat |
Informative Path Planning for Extreme Anomaly Detection in Environment
Exploration and Monitoring | An unmanned autonomous vehicle (UAV) is sent on a mission to explore and
reconstruct an unknown environment from a series of measurements collected by
Bayesian optimization. The success of the mission is judged by the UAV's
ability to faithfully reconstruct any anomalous features present in the
environment, with emphasis on the extremes (e.g., extreme topographic
depressions or abnormal chemical concentrations). We show that the criteria
commonly used for determining which locations the UAV should visit are
ill-suited for this task. We introduce a number of novel criteria that guide
the UAV towards regions of strong anomalies by leveraging previously collected
information in a mathematically elegant and computationally tractable manner.
We demonstrate superiority of the proposed approach in several applications,
including reconstruction of seafloor topography from real-world bathymetry
data, as well as tracking of dynamic anomalies. A particularly attractive
property of our approach is its ability to overcome adversarial conditions,
that is, situations in which prior beliefs about the locations of the extremes
are imprecise or erroneous.
| stat |
Context-Based Dynamic Pricing with Online Clustering | We consider a context-based dynamic pricing problem of online products which
have low sales. Sales data from Alibaba, a major global online retailer,
illustrate the prevalence of low-sale products. For these products, existing
single-product dynamic pricing algorithms do not work well due to insufficient
data samples. To address this challenge, we propose pricing policies that
concurrently perform clustering over products and set individual pricing
decisions on the fly. By clustering data and identifying products that have
similar demand patterns, we utilize sales data from products within the same
cluster to improve demand estimation and allow for better pricing decisions. We
evaluate the algorithms using the regret, and the result shows that when
product demand functions come from multiple clusters, our algorithms
significantly outperform traditional single-product pricing policies. Numerical
experiments using a real dataset from Alibaba demonstrate that the proposed
policies, compared with several benchmark policies, increase the revenue. The
results show that online clustering is an effective approach to tackling
dynamic pricing problems associated with low-sale products. Our algorithms were
further implemented in a field study at Alibaba with 40 products for 30
consecutive days, and compared to the products which use business-as-usual
pricing policy of Alibaba. The results from the field experiment show that the
overall revenue increased by 10.14%.
| stat |
Differential Privacy for Sparse Classification Learning | In this paper, we present a differential privacy version of convex and
nonconvex sparse classification approach. Based on alternating direction method
of multiplier (ADMM) algorithm, we transform the solving of sparse problem into
the multistep iteration process. Then we add exponential noise to stable steps
to achieve privacy protection. By the property of the post-processing holding
of differential privacy, the proposed approach satisfies the
$\epsilon-$differential privacy even when the original problem is unstable.
Furthermore, we present the theoretical privacy bound of the differential
privacy classification algorithm. Specifically, the privacy bound of our
algorithm is controlled by the algorithm iteration number, the privacy
parameter, the parameter of loss function, ADMM pre-selected parameter, and the
data size. Finally we apply our framework to logistic regression with $L_1$
regularizer and logistic regression with $L_{1/2}$ regularizer. Numerical
studies demonstrate that our method is both effective and efficient which
performs well in sensitive data analysis.
| stat |
Policy Learning with Adaptively Collected Data | Learning optimal policies from historical data enables the gains from
personalization to be realized in a wide variety of applications. The growing
policy learning literature focuses on a setting where the treatment assignment
policy does not adapt to the data. However, adaptive data collection is
becoming more common in practice, from two primary sources: 1) data collected
from adaptive experiments that are designed to improve inferential efficiency;
2) data collected from production systems that are adaptively evolving an
operational policy to improve performance over time (e.g. contextual bandits).
In this paper, we aim to address the challenge of learning the optimal policy
with adaptively collected data and provide one of the first theoretical
inquiries into this problem. We propose an algorithm based on generalized
augmented inverse propensity weighted estimators and establish its
finite-sample regret bound. We complement this regret upper bound with a lower
bound that characterizes the fundamental difficulty of policy learning with
adaptive data. Finally, we demonstrate our algorithm's effectiveness using both
synthetic data and public benchmark datasets.
| stat |
Early Detection of Long Term Evaluation Criteria in Online Controlled
Experiments | A common dilemma encountered by many upon implementing an optimization method
or experiment, whether it be a reinforcement learning algorithm, or A/B
testing, is deciding on what metric to optimize for. Very often short-term
metrics, which are easier to measure are chosen over long term metrics which
have undesirable time considerations and often a more complex calculation. In
this paper, we argue the importance of choosing a metrics that focuses on long
term effects. With this comes the necessity in the ability to measure
significant differences between groups relatively early. We present here an
efficient methodology for early detection of lifetime differences between
groups based on bootstrap hypothesis testing of the lifetime forecast of the
response. We present an application of this method in the domain of online
advertising and we argue that approach not only allows one to focus on the
ultimate metric of importance but also provides a means of accelerating the
testing period.
| stat |
Contextual Bandits with Continuous Actions: Smoothing, Zooming, and
Adapting | We study contextual bandit learning with an abstract policy class and
continuous action space. We obtain two qualitatively different regret bounds:
one competes with a smoothed version of the policy class under no continuity
assumptions, while the other requires standard Lipschitz assumptions. Both
bounds exhibit data-dependent "zooming" behavior and, with no tuning, yield
improved guarantees for benign problems. We also study adapting to unknown
smoothness parameters, establishing a price-of-adaptivity and deriving optimal
adaptive algorithms that require no additional information.
| stat |
What needles do sparse neural networks find in nonlinear haystacks | Using a sparsity inducing penalty in artificial neural networks (ANNs) avoids
over-fitting, especially in situations where noise is high and the training set
is small in comparison to the number of features. For linear models, such an
approach provably also recovers the important features with high probability in
regimes for a well-chosen penalty parameter. The typical way of setting the
penalty parameter is by splitting the data set and performing the
cross-validation, which is (1) computationally expensive and (2) not desirable
when the data set is already small to be further split (for example,
whole-genome sequence data). In this study, we establish the theoretical
foundation to select the penalty parameter without cross-validation based on
bounding with a high probability the infinite norm of the gradient of the loss
function at zero under the zero-feature assumption. Our approach is a
generalization of the universal threshold of Donoho and Johnstone (1994) to
nonlinear ANN learning. We perform a set of comprehensive Monte Carlo
simulations on a simple model, and the numerical results show the effectiveness
of the proposed approach.
| stat |
Deep imagination is a close to optimal policy for planning in large
decision trees under limited resources | Many decisions involve choosing an uncertain course of actions in deep and
wide decision trees, as when we plan to visit an exotic country for vacation.
In these cases, exhaustive search for the best sequence of actions is not
tractable due to the large number of possibilities and limited time or
computational resources available to make the decision. Therefore, planning
agents need to balance breadth (exploring many actions at each level of the
tree) and depth (exploring many levels in the tree) to allocate optimally their
finite search capacity. We provide efficient analytical solutions and numerical
analysis to the problem of allocating finite sampling capacity in one shot to
large decision trees. We find that in general the optimal policy is to allocate
few samples per level so that deep levels can be reached, thus favoring depth
over breadth search. In contrast, in poor environments and at low capacity, it
is best to broadly sample branches at the cost of not sampling deeply, although
this policy is marginally better than deep allocations. Our results provide a
theoretical foundation for the optimality of deep imagination for planning and
show that it is a generally valid heuristic that could have evolved from the
finite constraints of cognitive systems.
| stat |
Resolving learning rates adaptively by locating Stochastic Non-Negative
Associated Gradient Projection Points using line searches | Learning rates in stochastic neural network training are currently determined
a priori to training, using expensive manual or automated iterative tuning.
This study proposes gradient-only line searches to resolve the learning rate
for neural network training algorithms. Stochastic sub-sampling during training
decreases computational cost and allows the optimization algorithms to progress
over local minima. However, it also results in discontinuous cost functions.
Minimization line searches are not effective in this context, as they use a
vanishing derivative (first order optimality condition), which often do not
exist in a discontinuous cost function and therefore converge to
discontinuities as opposed to minima from the data trends. Instead, we base
candidate solutions along a search direction purely on gradient information, in
particular by a directional derivative sign change from negative to positive (a
Non-negative Associative Gradient Projection Point (NN- GPP)). Only considering
a sign change from negative to positive always indicates a minimum, thus
NN-GPPs contain second order information. Conversely, a vanishing gradient is
purely a first order condition, which may indicate a minimum, maximum or saddle
point. This insight allows the learning rate of an algorithm to be reliably
resolved as the step size along a search direction, increasing convergence
performance and eliminating an otherwise expensive hyperparameter.
| stat |
Optimal Rerandomization via a Criterion that Provides Insurance Against
Failed Experiments | We present an optimized rerandomization design procedure for a non-sequential
treatment-control experiment. Randomized experiments are the gold standard for
finding causal effects in nature. But sometimes random assignments result in
unequal partitions of the treatment and control group visibly seen as imbalance
in observed covariates. There can additionally be imbalance on unobserved
covariates. Imbalance in either observed or unobserved covariates increases
treatment effect estimator error inflating the width of confidence regions and
reducing experimental power. "Rerandomization" is a strategy that omits poor
imbalance assignments by limiting imbalance in the observed covariates to a
prespecified threshold. However, limiting this threshold too much can increase
the risk of contracting error from unobserved covariates. We introduce a
criterion that combines observed imbalance while factoring in the risk of
inadvertently imbalancing unobserved covariates. We then use this criterion to
locate the optimal rerandomization threshold based on the practitioner's level
of desired insurance against high estimator error. We demonstrate the gains of
our designs in simulation and in a dataset from a large randomized experiment
in education. We provide an open source R package available on CRAN named
OptimalRerandExpDesigns which generates designs according to our algorithm.
| stat |
A Scale-free Approach for False Discovery Rate Control in Generalized
Linear Models | The generalized linear models (GLM) have been widely used in practice to
model non-Gaussian response variables. When the number of explanatory features
is relatively large, scientific researchers are of interest to perform
controlled feature selection in order to simplify the downstream analysis. This
paper introduces a new framework for feature selection in GLMs that can achieve
false discovery rate (FDR) control in two asymptotic regimes. The key step is
to construct a mirror statistic to measure the importance of each feature,
which is based upon two (asymptotically) independent estimates of the
corresponding true coefficient obtained via either the data-splitting method or
the Gaussian mirror method. The FDR control is achieved by taking advantage of
the mirror statistic's property that, for any null feature, its sampling
distribution is (asymptotically) symmetric about 0. In the moderate-dimensional
setting in which the ratio between the dimension (number of features) p and the
sample size n converges to a fixed value, we construct the mirror statistic
based on the maximum likelihood estimation. In the high-dimensional setting
where p is much larger than n, we use the debiased Lasso to build the mirror
statistic. Compared to the Benjamini-Hochberg procedure, which crucially relies
on the asymptotic normality of the Z statistic, the proposed methodology is
scale free as it only hinges on the symmetric property, thus is expected to be
more robust in finite-sample cases. Both simulation results and a real data
application show that the proposed methods are capable of controlling the FDR,
and are often more powerful than existing methods including the
Benjamini-Hochberg procedure and the knockoff filter.
| stat |
Scalable and Accurate Variational Bayes for High-Dimensional Binary
Regression Models | Modern methods for Bayesian regression beyond the Gaussian response setting
are computationally impractical or inaccurate in high dimensions. As discussed
in recent literature, bypassing this trade-off is still an open problem even in
basic binary regression models, and there is limited theory on the quality of
variational approximations in high-dimensional settings. To address this gap,
we study the approximation accuracy of routine-use mean-field variational Bayes
in high-dimensional probit regression with Gaussian priors, obtaining new and
practically relevant results on the pathological behavior of this strategy in
uncertainty quantification, estimation and prediction, that also suggest
caution against maximum a posteriori estimates when p>n. Motivated by these
results, we develop a new partially-factorized variational approximation for
the posterior distribution of the probit coefficients that leverages a
representation with global and local variables but, unlike for classical
mean-field assumptions, it avoids a fully factorized approximation, and instead
assumes a factorization only for local variables. We prove that the resulting
approximation belongs to a tractable class of unified skew-normal densities
that incorporates skewness and, unlike for state-of-the-art mean-field
solutions, converges to the exact posterior density as p goes to infinity. To
solve the variational optimization problem, we derive a tractable CAVI
algorithm that easily scales to p in tens of thousands, and provably requires a
number of iterations converging to 1 as p goes to infinity. Such findings are
also illustrated in extensive empirical studies where our new solution is shown
to improve the accuracy of mean-field variational Bayes for any n and p, with
the magnitude of these gains being remarkable in those high-dimensional p>n
settings where state-of-the-art methods are computationally impractical.
| stat |
A Competing Risks Model with Binary Time Varying Covariates for
Estimation of Breast Cancer Risks in BRCA1 Families | Mammographic screening and prophylactic surgery such as risk-reducing
salpingo oophorectomy (RRSO) can potentially reduce breast cancer risks among
mutation carriers of BRCA families. The evaluation of these interventions is
usually complicated by the fact that their effects on breast cancer may change
over time and by the presence of competing risks. We introduce a correlated
competing risks model to model breast and ovarian cancer risks within BRCA1
families that accounts for time-varying covariates (TVCs). Different parametric
forms for these TVCs are proposed for more flexibility and a correlated gamma
frailty model is specified to account for the correlated competing events. We
also introduced a new ascertainment correction approach that accounts for the
selection of families through probands affected with either breast or ovarian
cancer, or unaffected. Our simulation studies demonstrate the good performances
of our proposed approach in terms of bias and precision of the estimators of
model parameters and cause-specific penetrances over different levels of
familial correlations. We apply our new approach to 498 BRCA1 mutation carrier
families recruited through the Breast Cancer Family Registry. Our results
demonstrate the importance of the functional form of the TVC when assessing the
role of RRSO on breast cancer. In particular, under the best fitting TVC model,
the overall effect of RRSO on breast cancer risk was statistically significant
in women with BRCA1.
| stat |
Snowboot: Bootstrap Methods for Network Inference | Complex networks are used to describe a broad range of disparate social
systems and natural phenomena, from power grids to customer segmentation to
human brain connectome. Challenges of parametric model specification and
validation inspire a search for more data-driven and flexible nonparametric
approaches for inference of complex networks. In this paper we discuss
methodology and R implementation of two bootstrap procedures on random
networks, that is, patchwork bootstrap of Thompson et al. (2016) and Gel et al.
(2017) and vertex bootstrap of Snijders and Borgatti (1999). To our knowledge,
the new R package snowboot is the first implementation of the vertex and
patchwork bootstrap inference on networks in R. Our new package is accompanied
with a detailed user's manual, and is compatible with the popular R package on
network studies igraph. We evaluate the patchwork bootstrap and vertex
bootstrap with extensive simulation studies and illustrate their utility in
application to analysis of real world networks.
| stat |
Short- and long-term forecasting of electricity prices using embedding
of calendar information in neural networks | Electricity prices strongly depend on seasonality on different time scales,
therefore any forecasting of electricity prices has to account for it. Neural
networks proved successful in forecasting, but complicated architectures like
LSTM are used to integrate the seasonal behavior. This paper shows that simple
neural networks architectures like DNNs with an embedding layer for seasonality
information deliver not only a competitive but superior forecast. The embedding
based processing of calendar information additionally opens up new applications
for neural networks in electricity trading like the generation of price forward
curves. Besides the theoretical foundation, this paper also provides an
empirical multi-year study on the German electricity market for both
applications and derives economical insights from the embedding layer. The
study shows that in short-term price-forecasting the mean absolute error of the
proposed neural networks with embedding layer is only about half of the mean
absolute forecast error of state-of-the-art LSTM approaches. The predominance
of the proposed approach is also supported by a statistical analysis using
Friedman and Holm's tests.
| stat |
Linear shrinkage for predicting responses in large-scale multivariate
linear regression | We propose a new prediction method for multivariate linear regression
problems where the number of features is less than the sample size but the
number of outcomes is extremely large. Many popular procedures, such as
penalized regression procedures, require parameter tuning that is
computationally untenable in such large-scale problems. We take a different
approach, motivated by ideas from simultaneous estimation problems, that
performs linear shrinkage on ordinary least squares parameter estimates. Our
approach is extremely computationally efficient and tuning-free. We show that
it can asymptotically outperform ordinary least squares without any structural
assumptions on the true regression coefficients and illustrate its good
performance in simulations and an analysis of single-cell RNA-seq data.
| stat |
A Tree-based Federated Learning Approach for Personalized Treatment
Effect Estimation from Heterogeneous Data Sources | Federated learning is an appealing framework for analyzing sensitive data
from distributed health data networks. Under this framework, data partners at
local sites collaboratively build an analytical model under the orchestration
of a coordinating site, while keeping the data decentralized. While integrating
information from multiple sources may boost statistical efficiency, existing
federated learning methods mainly assume data across sites are homogeneous
samples of the global population, failing to properly account for the extra
variability across sites in estimation and inference. Drawing on a
multi-hospital electronic health records network, we develop an efficient and
interpretable tree-based ensemble of personalized treatment effect estimators
to join results across hospital sites, while actively modeling for the
heterogeneity in data sources through site partitioning. The efficiency of this
approach is demonstrated by a study of causal effects of oxygen saturation on
hospital mortality and backed up by comprehensive numerical results.
| stat |
GARCH-UGH: A bias-reduced approach for dynamic extreme Value-at-Risk
estimation in financial time series | The Value-at-Risk (VaR) is a widely used instrument in financial risk
management. The question of estimating the VaR of loss return distributions at
extreme levels is an important question in financial applications, both from
operational and regulatory perspectives; in particular, the dynamic estimation
of extreme VaR given the recent past has received substantial attention. We
propose here a two-step bias-reduced estimation methodology called GARCH-UGH
(Unbiased Gomes-de Haan), whereby financial returns are first filtered using an
AR-GARCH model, and then a bias-reduced estimator of extreme quantiles is
applied to the standardized residuals to estimate one-step ahead dynamic
extreme VaR. Our results indicate that the GARCH-UGH estimates are more
accurate than those obtained by combining conventional AR-GARCH filtering and
extreme value estimates from the perspective of in-sample and out-of-sample
backtestings of historical daily returns on several financial time series.
| stat |
Opening practice: supporting Reproducibility and Critical spatial data
science | This paper reflects on a number of trends towards a more open and
reproducible approach to geographic and spatial data science over recent years.
In particular it considers trends towards Big Data, and the impacts this is
having on spatial data analysis and modelling. It identifies a turn in academia
towards coding as a core analytic tool, and away from proprietary software
tools offering 'black boxes' where the internal workings of the analysis are
not revealed. It is argued that this closed form software is problematic, and
considers a number of ways in which issues identified in spatial data analysis
(such as the MAUP) could be overlooked when working with closed tools, leading
to problems of interpretation and possibly inappropriate actions and policies
based on these. In addition, this paper and considers the role that
reproducible and open spatial science may play in such an approach, taking into
account the issues raised. It highlights the dangers of failing to account for
the geographical properties of data, now that all data are spatial (they are
collected somewhere), the problems of a desire for n=all observations in data
science and it identifies the need for a critical approach. This is one in
which openness, transparency, sharing and reproducibility provide a mantra for
defensible and robust spatial data science.
| stat |
Survival trees for right-censored data based on score based parameter
instability test | Survival analysis of right censored data arises often in many areas of
research including medical research. Effect of covariates (and their
interactions) on survival distribution can be studied through existing methods
which requires to pre-specify the functional form of the covariates including
their interactions. Survival trees offer relatively flexible approach when the
form of covariates' effects is unknown. Most of the currently available
survival tree construction techniques are not based on a formal test of
significance; however, recently proposed ctree algorithm (Hothorn et al., 2006)
uses permutation test for splitting decision that may be conservative at times.
We consider parameter instability test of statistical significance of
heterogeneity to guard against spurious findings of variation in covariates'
effect without being overly conservative. We have proposed SurvCART algorithm
to construct survival tree under conditional inference framework (Hothorn et
al., 2006) that selects splitting variable via parameter instability test and
subsequently finds the optimal split based on some maximally chosen statistic.
Notably, unlike the existing algorithms which focuses only on heterogeneity in
event time distribution, the proposed SurvCART algorithm can take splitting
decision based in censoring distribution as well along with heterogeneity in
event time distribution. The operating characteristics of parameter instability
test and comparative assessment of SurvCART algorithm were carried out via
simulation. Finally, SurvCART algorithm was applied to a real data setting. The
proposed method is fully implemented in R package LongCART available on CRAN.
| stat |
Bayesian Screening: Multi-test Bayesian Optimization Applied to in
silico Material Screening | We present new multi-test Bayesian optimization models and algorithms for use
in large scale material screening applications. Our screening problems are
designed around two tests, one expensive and one cheap. This paper differs from
other recent work on multi-test Bayesian optimization through use of a flexible
model that allows for complex, non-linear relationships between the cheap and
expensive test scores. This additional modeling flexibility is essential in the
material screening applications which we describe. We demonstrate the power of
our new algorithms on a family of synthetic toy problems as well as on real
data from two large scale screening studies.
| stat |
Response-adaptive randomization in clinical trials: from myths to
practical considerations | Response-adaptive randomization (RAR) is part of a wider class of
data-dependent sampling algorithms, for which clinical trials have commonly
been used as a motivating application. In that context, patient allocation to
treatments is defined using the accrued data on responses to alter
randomization probabilities, in order to achieve different experimental goals.
RAR has received abundant theoretical attention from the biostatistical
literature since the 1930's and has been the subject of heated debates.
Recently it has received renewed consideration from the applied community due
to successful practical examples and its widespread use in machine learning.
Many position papers on the subject present a one-sided view on its use, which
is of limited value for the non-expert. This work aims to address this
persistent gap by providing a critical, balanced and updated review of
methodological and practical issues to consider when debating the use of RAR in
clinical trials.
| stat |
An exact solver for the Weston-Watkins SVM subproblem | Recent empirical evidence suggests that the Weston-Watkins support vector
machine is among the best performing multiclass extensions of the binary SVM.
Current state-of-the-art solvers repeatedly solve a particular subproblem
approximately using an iterative strategy. In this work, we propose an
algorithm that solves the subproblem exactly using a novel reparametrization of
the Weston-Watkins dual problem. For linear WW-SVMs, our solver shows
significant speed-up over the state-of-the-art solver when the number of
classes is large. Our exact subproblem solver also allows us to prove linear
convergence of the overall solver.
| stat |
Deep Factors for Forecasting | Producing probabilistic forecasts for large collections of similar and/or
dependent time series is a practically relevant and challenging task. Classical
time series models fail to capture complex patterns in the data, and
multivariate techniques struggle to scale to large problem sizes. Their
reliance on strong structural assumptions makes them data-efficient, and allows
them to provide uncertainty estimates. The converse is true for models based on
deep neural networks, which can learn complex patterns and dependencies given
enough data. In this paper, we propose a hybrid model that incorporates the
benefits of both approaches. Our new method is data-driven and scalable via a
latent, global, deep component. It also handles uncertainty through a local
classical model. We provide both theoretical and empirical evidence for the
soundness of our approach through a necessary and sufficient decomposition of
exchangeable time series into a global and a local part. Our experiments
demonstrate the advantages of our model both in term of data efficiency,
accuracy and computational complexity.
| stat |
A Square-Root Second-Order Extended Kalman Filtering Approach for
Estimating Smoothly Time-Varying Parameters | Researchers collecting intensive longitudinal data (ILD) are increasingly
looking to model psychological processes, such as emotional dynamics, that
organize and adapt across time in complex and meaningful ways. This is also the
case for researchers looking to characterize the impact of an intervention on
individual behavior. To be useful, statistical models must be capable of
characterizing these processes as complex, time-dependent phenomenon, otherwise
only a fraction of the system dynamics will be recovered. In this paper we
introduce a Square-Root Second-Order Extended Kalman Filtering approach for
estimating smoothly time-varying parameters. This approach is capable of
handling dynamic factor models where the relations between variables underlying
the processes of interest change in a manner that may be difficult to specify
in advance. We examine the performance of our approach in a Monte Carlo
simulation and show the proposed algorithm accurately recovers the unobserved
states in the case of a bivariate dynamic factor model with time-varying
dynamics and treatment effects. Furthermore, we illustrate the utility of our
approach in characterizing the time-varying effect of a meditation intervention
on day-to-day emotional experiences.
| stat |
A Gibbs sampler for a class of random convex polytopes | We present a Gibbs sampler for the Dempster-Shafer (DS) approach to
statistical inference for Categorical distributions. The DS framework extends
the Bayesian approach, allows in particular the use of partial prior
information, and yields three-valued uncertainty assessments representing
probabilities "for", "against", and "don't know" about formal assertions of
interest. The proposed algorithm targets the distribution of a class of random
convex polytopes which encapsulate the DS inference. The sampler relies on an
equivalence between the iterative constraints of the vertex configuration and
the non-negativity of cycles in a fully connected directed graph. Illustrations
include the testing of independence in 2x2 contingency tables and parameter
estimation of the linkage model.
| stat |
A Common Atom Model for the Bayesian Nonparametric Analysis of Nested
Data | The use of high-dimensional data for targeted therapeutic interventions
requires new ways to characterize the heterogeneity observed across subgroups
of a specific population. In particular, models for partially exchangeable data
are needed for inference on nested datasets, where the observations are assumed
to be organized in different units and some sharing of information is required
to learn distinctive features of the units. In this manuscript, we propose a
nested Common Atoms Model (CAM) that is particularly suited for the analysis of
nested datasets where the distributions of the units are expected to differ
only over a small fraction of the observations sampled from each unit. The
proposed CAM allows a two-layered clustering at the distributional and
observational level and is amenable to scalable posterior inference through the
use of a computationally efficient nested slice-sampler algorithm. We further
discuss how to extend the proposed modeling framework to handle discrete
measurements, and we conduct posterior inference on a real microbiome dataset
from a diet swap study to investigate how the alterations in intestinal
microbiota composition are associated with different eating habits. We further
investigate the performance of our model in capturing true distributional
structures in the population by means of a simulation study.
| stat |
Traversing the noise of dynamic mini-batch sub-sampled loss functions: A
visual guide | Mini-batch sub-sampling in neural network training is unavoidable, due to
growing data demands, memory-limited computational resources such as graphical
processing units (GPUs), and the dynamics of on-line learning. In this study we
specifically distinguish between static mini-batch sub-sampled loss functions,
where mini-batches are intermittently fixed during training, resulting in
smooth but biased loss functions; and the dynamic sub-sampling equivalent,
where new mini-batches are sampled at every loss evaluation, trading bias for
variance in sampling induced discontinuities. These render automated
optimization strategies such as minimization line searches ineffective, since
critical points may not exist and function minimizers find spurious,
discontinuity induced minima.
This paper suggests recasting the optimization problem to find stochastic
non-negative associated gradient projection points (SNN-GPPs). We demonstrate
that the SNN-GPP optimality criterion is less susceptible to sub-sampling
induced discontinuities than critical points or minimizers. We conduct a visual
investigation, comparing local minimum and SNN-GPP optimality criteria in the
loss functions of a simple neural network training problem for a variety of
popular activation functions. Since SNN-GPPs better approximate the location of
true optima, particularly when using smooth activation functions with high
curvature characteristics, we postulate that line searches locating SNN-GPPs
can contribute significantly to automating neural network training
| stat |
Nearest Neighbor and Kernel Survival Analysis: Nonasymptotic Error
Bounds and Strong Consistency Rates | We establish the first nonasymptotic error bounds for Kaplan-Meier-based
nearest neighbor and kernel survival probability estimators where feature
vectors reside in metric spaces. Our bounds imply rates of strong consistency
for these nonparametric estimators and, up to a log factor, match an existing
lower bound for conditional CDF estimation. Our proof strategy also yields
nonasymptotic guarantees for nearest neighbor and kernel variants of the
Nelson-Aalen cumulative hazards estimator. We experimentally compare these
methods on four datasets. We find that for the kernel survival estimator, a
good choice of kernel is one learned using random survival forests.
| stat |
Short-term forecasting of solar irradiance without local telemetry: a
generalized model using satellite data | Due to the increasing integration of solar power into the electrical grid,
forecasting short-term solar irradiance has become key for many applications,
e.g.~operational planning, power purchases, reserve activation, etc. In this
context, as solar generators are geographically dispersed and ground
measurements are not always easy to obtain, it is very important to have
general models that can predict solar irradiance without the need of local
data. In this paper, a model that can perform short-term forecasting of solar
irradiance in any general location without the need of ground measurements is
proposed. To do so, the model considers satellite-based measurements and
weather-based forecasts, and employs a deep neural network structure that is
able to generalize across locations; particularly, the network is trained only
using a small subset of sites where ground data is available, and the model is
able to generalize to a much larger number of locations where ground data does
not exist. As a case study, 25 locations in The Netherlands are considered and
the proposed model is compared against four local models that are individually
trained for each location using ground measurements. Despite the general nature
of the model, it is shown show that the proposed model is equal or better than
the local models: when comparing the average performance across all the
locations and prediction horizons, the proposed model obtains a 31.31% rRMSE
(relative root mean square error) while the best local model achieves a 32.01%
rRMSE.
| stat |
Angle-Based Cost-Sensitive Multicategory Classification | Many real-world classification problems come with costs which can vary for
different types of misclassification. It is thus important to develop
cost-sensitive classifiers which minimize the total misclassification cost.
Although binary cost-sensitive classifiers have been well-studied, solving
multicategory classification problems is still challenging. A popular approach
to address this issue is to construct K classification functions for a K-class
problem and remove the redundancy by imposing a sum-to-zero constraint.
However, such method usually results in higher computational complexity and
inefficient algorithms. In this paper, we propose a novel angle-based
cost-sensitive classification framework for multicategory classification
without the sum-to-zero constraint. Loss functions that included in the
angle-based cost-sensitive classification framework are further justified to be
Fisher consistent. To show the usefulness of the framework, two cost-sensitive
multicategory boosting algorithms are derived as concrete instances. Numerical
experiments demonstrate that proposed boosting algorithms yield competitive
classification performances against other existing boosting approaches.
| stat |
Sequential Bayesian optimal experimental design for structural
reliability analysis | Structural reliability analysis is concerned with estimation of the
probability of a critical event taking place, described by $P(g(\textbf{X})
\leq 0)$ for some $n$-dimensional random variable $\textbf{X}$ and some
real-valued function $g$. In many applications the function $g$ is practically
unknown, as function evaluation involves time consuming numerical simulation or
some other form of experiment that is expensive to perform. The problem we
address in this paper is how to optimally design experiments, in a Bayesian
decision theoretic fashion, when the goal is to estimate the probability
$P(g(\textbf{X}) \leq 0)$ using a minimal amount of resources. As opposed to
existing methods that have been proposed for this purpose, we consider a
general structural reliability model given in hierarchical form. We therefore
introduce a general formulation of the experimental design problem, where we
distinguish between the uncertainty related to the random variable $\textbf{X}$
and any additional epistemic uncertainty that we want to reduce through
experimentation. The effectiveness of a design strategy is evaluated through a
measure of residual uncertainty, and efficient approximation of this quantity
is crucial if we want to apply algorithms that search for an optimal strategy.
The method we propose is based on importance sampling combined with the
unscented transform for epistemic uncertainty propagation. We implement this
for the myopic (one-step look ahead) alternative, and demonstrate the
effectiveness through a series of numerical experiments.
| stat |
Toroidal diffusions and protein structure evolution | This chapter shows how toroidal diffusions are convenient methodological
tools for modelling protein evolution in a probabilistic framework. The chapter
addresses the construction of ergodic diffusions with stationary distributions
equal to well-known directional distributions, which can be regarded as
toroidal analogues of the Ornstein-Uhlenbeck process. The important challenges
that arise in the estimation of the diffusion parameters require the
consideration of tractable approximate likelihoods and, among the several
approaches introduced, the one yielding a specific approximation to the
transition density of the wrapped normal process is shown to give the best
empirical performance on average. This provides the methodological building
block for Evolutionary Torus Dynamic Bayesian Network (ETDBN), a hidden Markov
model for protein evolution that emits a wrapped normal process and two
continuous-time Markov chains per hidden state. The chapter describes the main
features of ETDBN, which allows for both "smooth" conformational changes and
"catastrophic" conformational jumps, and several empirical benchmarks. The
insights into the relationship between sequence and structure evolution that
ETDBN provides are illustrated in a case study.
| stat |
Mortality Forecasting using Factor Models: Time-varying or
Time-invariant Factor Loadings? | Many existing mortality models follow the framework of classical factor
models, such as the Lee-Carter model and its variants. Latent common factors in
factor models are defined as time-related mortality indices (such as $\kappa_t$
in the Lee-Carter model). Factor loadings, which capture the linear
relationship between age variables and latent common factors (such as $\beta_x$
in the Lee-Carter model), are assumed to be time-invariant in the classical
framework. This assumption is usually too restrictive in reality as mortality
datasets typically span a long period of time. Driving forces such as medical
improvement of certain diseases, environmental changes and technological
progress may significantly influence the relationship of different variables.
In this paper, we first develop a factor model with time-varying factor
loadings (time-varying factor model) as an extension of the classical factor
model for mortality modelling. Two forecasting methods to extrapolate the
factor loadings, the local regression method and the naive method, are proposed
for the time-varying factor model. From the empirical data analysis, we find
that the new model can capture the empirical feature of time-varying factor
loadings and improve mortality forecasting over different horizons and
countries. Further, we propose a novel approach based on change point analysis
to estimate the optimal `boundary' between short-term and long-term
forecasting, which is favoured by the local linear regression and naive method,
respectively. Additionally, simulation studies are provided to show the
performance of the time-varying factor model under various scenarios.
| stat |
Quantifying Sufficient Randomness for Causal Inference | Spurious association arises from covariance between propensity for the
treatment and individual risk for the outcome. For sensitivity analysis with
stochastic counterfactuals we introduce a methodology to characterize
uncertainty in causal inference from natural experiments and quasi-experiments.
Our sensitivity parameters are standardized measures of variation in propensity
and individual risk, and one minus their geometric mean is an intuitive measure
of randomness in the data generating process. Within our latent propensity-risk
model, we show how to compute from contingency table data a threshold, $T$, of
sufficient randomness for causal inference. If the actual randomness of the
data generating process exceeds this threshold then causal inference is
warranted.
| stat |
Generalized reliability based on distances | The intraclass correlation coefficient (ICC) is a classical index of
measurement reliability. With the advent of new and complex types of data for
which the ICC is not defined, there is a need for new ways to assess
reliability. To meet this need, we propose a new distance-based intraclass
correlation coefficient (dbICC), defined in terms of arbitrary distances among
observations. We introduce a bias correction to improve the coverage of
bootstrap confidence intervals for the dbICC, and demonstrate its efficacy via
simulation. We illustrate the proposed method by analyzing the test-retest
reliability of brain connectivity matrices derived from a set of repeated
functional magnetic resonance imaging scans. The Spearman-Brown formula, which
shows how more intensive measurement increases reliability, is extended to
encompass the dbICC.
| stat |
Discontinuous Hamiltonian Monte Carlo for discrete parameters and
discontinuous likelihoods | Hamiltonian Monte Carlo has emerged as a standard tool for posterior
computation. In this article, we present an extension that can efficiently
explore target distributions with discontinuous densities. Our extension in
particular enables efficient sampling from ordinal parameters though embedding
of probability mass functions into continuous spaces. We motivate our approach
through a theory of discontinuous Hamiltonian dynamics and develop a
corresponding numerical solver. The proposed solver is the first of its kind,
with a remarkable ability to exactly preserve the Hamiltonian. We apply our
algorithm to challenging posterior inference problems to demonstrate its wide
applicability and competitive performance.
| stat |
Sparse semiparametric canonical correlation analysis for data of mixed
types | Canonical correlation analysis investigates linear relationships between two
sets of variables, but often works poorly on modern data sets due to
high-dimensionality and mixed data types such as continuous, binary and
zero-inflated. To overcome these challenges, we propose a semiparametric
approach for sparse canonical correlation analysis based on Gaussian copula.
Our main contribution is a truncated latent Gaussian copula model for data with
excess zeros, which allows us to derive a rank-based estimator of the latent
correlation matrix for mixed variable types without the estimation of marginal
transformation functions. The resulting canonical correlation analysis method
works well in high-dimensional settings as demonstrated via numerical studies,
as well as in application to the analysis of association between gene
expression and micro RNA data of breast cancer patients.
| stat |
The Reciprocal Bayesian LASSO | A reciprocal LASSO (rLASSO) regularization employs a decreasing penalty
function as opposed to conventional penalization methods that use increasing
penalties on the coefficients, leading to stronger parsimony and superior model
selection relative to traditional shrinkage methods. Here we consider a fully
Bayesian formulation of the rLASSO problem, which is based on the observation
that the rLASSO estimate for linear regression parameters can be interpreted as
a Bayesian posterior mode estimate when the regression parameters are assigned
independent inverse Laplace priors. Bayesian inference from this posterior is
possible using an expanded hierarchy motivated by a scale mixture of double
Pareto or truncated normal distributions. On simulated and real datasets, we
show that the Bayesian formulation outperforms its classical cousin in
estimation, prediction, and variable selection across a wide range of scenarios
while offering the advantage of posterior inference. Finally, we discuss other
variants of this new approach and provide a unified framework for variable
selection using flexible reciprocal penalties. All methods described in this
paper are publicly available as an R package at:
https://github.com/himelmallick/BayesRecipe.
| stat |
Understanding the 2016 US Presidential Election using ecological
inference and distribution regression with census microdata | We combine fine-grained spatially referenced census data with the vote
outcomes from the 2016 US presidential election. Using this dataset, we perform
ecological inference using distribution regression (Flaxman et al, KDD 2015)
with a multinomial-logit regression so as to model the vote outcome Trump,
Clinton, Other / Didn't vote as a function of demographic and socioeconomic
features. Ecological inference allows us to estimate "exit poll" style results
like what was Trump's support among white women, but for entirely novel
categories. We also perform exploratory data analysis to understand which
census variables are predictive of voting for Trump, voting for Clinton, or not
voting for either. All of our methods are implemented in Python and R, and are
available online for replication.
| stat |
Assessing the accuracy of record linkages with Markov chain based Monte
Carlo simulation approach | Record linkage is the process of finding matches and linking records from
different data sources so that the linked records belong to the same entity.
There is an increasing number of applications of record linkage in statistical,
health, government and business organisations to link administrative, survey,
population census and other files to create a complete set of information for
more complete and comprehensive analysis. To make valid inferences using a
linked file, it is increasingly becoming important to assess the linking
method. It is also important to find techniques to improve the linking process
to achieve higher accuracy. This motivates to develop a method for assessing
linking process and help decide which linking method is likely to be more
accurate for a linking task. This paper proposes a Markov Chain based Monte
Carlo simulation approach, MaCSim for assessing a linking method and
illustrates the utility of the approach using a realistic synthetic dataset
received from the Australian Bureau of Statistics to avoid privacy issues
associated with using real personal information. A linking method applied by
MaCSim is also defined. To assess the defined linking method, correct re-link
proportions for each record are calculated using our developed simulation
approach. The accuracy is determined for a number of simulated datasets. The
analyses indicated promising performance of the proposed method MaCSim of the
assessment of accuracy of the linkages. The computational aspects of the
methodology are also investigated to assess its feasibility for practical use.
| stat |
A Geospatial Functional Model For OCO-2 Data with Application on
Imputation and Land Fraction Estimation | Data from NASA's Orbiting Carbon Observatory-2 (OCO-2) satellite is essential
to many carbon management strategies. A retrieval algorithm is used to estimate
CO2 concentration using the radiance data measured by OCO-2. However, due to
factors such as cloud cover and cosmic rays, the spatial coverage of the
retrieval algorithm is limited in some areas of critical importance for carbon
cycle science. Mixed land/water pixels along the coastline are also not used in
the retrieval processing due to the lack of valid ancillary variables including
land fraction. We propose an approach to model spatial spectral data to solve
these two problems by radiance imputation and land fraction estimation. The
spectral observations are modeled as spatially indexed functional data with
footprint-specific parameters and are reduced to much lower dimensions by
functional principal component analysis. The principal component scores are
modeled as random fields to account for the spatial dependence, and the missing
spectral observations are imputed by kriging the principal component scores.
The proposed method is shown to impute spectral radiance with high accuracy for
observations over the Pacific Ocean. An unmixing approach based on this model
provides much more accurate land fraction estimates in our validation study
along Greece coastlines.
| stat |
Revisiting the Train Loss: an Efficient Performance Estimator for Neural
Architecture Search | Reliable yet efficient evaluation of generalisation performance of a proposed
architecture is crucial to the success of neural architecture search (NAS).
Traditional approaches face a variety of limitations: training each
architecture to completion is prohibitively expensive, early stopping estimates
may correlate poorly with fully trained performance, and model-based estimators
require large training sets. Instead, motivated by recent results linking
training speed and generalisation with stochastic gradient descent, we propose
to estimate the final test performance based on the sum of training losses. Our
estimator is inspired by the marginal likelihood, which is used for Bayesian
model selection. Our model-free estimator is simple, efficient, and cheap to
implement, and does not require hyperparameter-tuning or surrogate training
before deployment. We demonstrate empirically that our estimator consistently
outperforms other baselines and can achieve a rank correlation of 0.95 with
final test accuracy on the NAS-Bench201 dataset within 50 epochs.
| stat |
An Early Stopping Bayesian Data Assimilation Approach for Mixed-Logit
Estimation | The mixed-logit model is a flexible tool in transportation choice analysis,
which provides valuable insights into inter and intra-individual behavioural
heterogeneity. However, applications of mixed-logit models are limited by the
high computational and data requirements for model estimation. When estimating
on small samples, the Bayesian estimation approach becomes vulnerable to over
and under-fitting. This is problematic for investigating the behaviour of
specific population sub-groups or market segments with low data availability.
Similar challenges arise when transferring an existing model to a new location
or time period, e.g., when estimating post-pandemic travel behaviour. We
propose an Early Stopping Bayesian Data Assimilation (ESBDA) simulator for
estimation of mixed-logit which combines a Bayesian statistical approach with
Machine Learning methodologies. The aim is to improve the transferability of
mixed-logit models and to enable the estimation of robust choice models with
low data availability. This approach can provide new insights into choice
behaviour where the traditional estimation of mixed-logit models was not
possible due to low data availability, and open up new opportunities for
investment and planning decisions support. The ESBDA estimator is benchmarked
against the Direct Application approach, a basic Bayesian simulator with random
starting parameter values and a Bayesian Data Assimilation (BDA) simulator
without early stopping. The ESBDA approach is found to effectively overcome
under and over-fitting and non-convergence issues in simulation. Its resulting
models clearly outperform those of the reference simulators in predictive
accuracy. Furthermore, models estimated with ESBDA tend to be more robust, with
significant parameters with signs and values consistent with behavioural
theory, even when estimated on small samples.
| stat |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.