title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
At the Intersection of Deep Sequential Model Framework and State-space
Model Framework: Study on Option Pricing | Inference and forecast problems of the nonlinear dynamical system have arisen
in a variety of contexts. Reservoir computing and deep sequential models, on
the one hand, have demonstrated efficient, robust, and superior performance in
modeling simple and chaotic dynamical systems. However, their innate
deterministic feature has partially detracted their robustness to noisy system,
and their inability to offer uncertainty measurement has also been an
insufficiency of the framework. On the other hand, the traditional state-space
model framework is robust to noise. It also carries measured uncertainty,
forming a just-right complement to the reservoir computing and deep sequential
model framework. We propose the unscented reservoir smoother, a model that
unifies both deep sequential and state-space models to achieve both frameworks'
superiorities. Evaluated in the option pricing setting on top of noisy
datasets, URS strikes highly competitive forecasting accuracy, especially those
of longer-term, and uncertainty measurement. Further extensions and
implications on URS are also discussed to generalize a full integration of both
frameworks.
| stat |
Integrating data science ethics into an undergraduate major | We present a programmatic approach to incorporating ethics into an
undergraduate major in statistical and data sciences. We discuss
departmental-level initiatives designed to meet the National Academy of
Sciences recommendation for weaving ethics into the curriculum from
top-to-bottom as our majors progress from our introductory courses to our
senior capstone course, as well as from side-to-side through co-curricular
programming. We also provide six examples of data science ethics modules used
in five different courses at our liberal arts college, each focusing on a
different ethical consideration. The modules are designed to be portable such
that they can be flexibly incorporated into existing courses at different
levels of instruction with minimal disruption to syllabi. We conclude with next
steps and preliminary assessments.
| stat |
Element-centric clustering comparison unifies overlaps and hierarchy | Clustering is one of the most universal approaches for understanding complex
data. A pivotal aspect of clustering analysis is quantitatively comparing
clusterings; clustering comparison is the basis for many tasks such as
clustering evaluation, consensus clustering, and tracking the temporal
evolution of clusters. In particular, the extrinsic evaluation of clustering
methods requires comparing the uncovered clusterings to planted clusterings or
known metadata. Yet, as we demonstrate, existing clustering comparison measures
have critical biases which undermine their usefulness, and no measure
accommodates both overlapping and hierarchical clusterings. Here we unify the
comparison of disjoint, overlapping, and hierarchically structured clusterings
by proposing a new element-centric framework: elements are compared based on
the relationships induced by the cluster structure, as opposed to the
traditional cluster-centric philosophy. We demonstrate that, in contrast to
standard clustering similarity measures, our framework does not suffer from
critical biases and naturally provides unique insights into how the clusterings
differ. We illustrate the strengths of our framework by revealing new insights
into the organization of clusters in two applications: the improved
classification of schizophrenia based on the overlapping and hierarchical
community structure of fMRI brain networks, and the disentanglement of various
social homophily factors in Facebook social networks. The universality of
clustering suggests far-reaching impact of our framework throughout all areas
of science.
| stat |
An adaptive stochastic optimization algorithm for resource allocation | We consider the classical problem of sequential resource allocation where a
decision maker must repeatedly divide a budget between several resources, each
with diminishing returns. This can be recast as a specific stochastic
optimization problem where the objective is to maximize the cumulative reward,
or equivalently to minimize the regret. We construct an algorithm that is {\em
adaptive} to the complexity of the problem, expressed in term of the regularity
of the returns of the resources, measured by the exponent in the {\L}ojasiewicz
inequality (or by their universal concavity parameter). Our
parameter-independent algorithm recovers the optimal rates for strongly-concave
functions and the classical fast rates of multi-armed bandit (for linear reward
functions). Moreover, the algorithm improves existing results on stochastic
optimization in this regret minimization setting for intermediate cases.
| stat |
A flexible model-free prediction-based framework for feature ranking | Despite the availability of numerous statistical and machine learning tools
for joint feature modeling, many scientists investigate features marginally,
i.e., one feature at a time. This is partly due to training and convention but
also roots in scientists' strong interests in simple visualization and
interpretability. As such, marginal feature ranking for some predictive tasks,
e.g., prediction of cancer driver genes, is widely practiced in the process of
scientific discoveries. In this work, we focus on marginal ranking for binary
prediction, the arguably most common predictive tasks. We argue that the most
widely used marginal ranking criteria, including the Pearson correlation, the
two-sample t test, and two-sample Wilcoxon rank-sum test, do not fully take
feature distributions and prediction objectives into account. To address this
gap in practice, we propose two ranking criteria corresponding to two
prediction objectives: the classical criterion (CC) and the Neyman-Pearson
criterion (NPC), both of which use model-free nonparametric implementations to
accommodate diverse feature distributions. Theoretically, we show that under
regularity conditions both criteria achieve sample-level ranking consistent
with their population-level counterpart with high probability. Moreover, NPC is
robust to sampling bias when the two class proportions in a sample deviate from
those in the population. This property endows NPC good potential in biomedical
research where sampling bias is common. We demonstrate the use and relative
advantages of CC and NPC in simulation and real data studies. Our model-free
objective-based ranking idea is extendable to ranking feature subsets and
generalizable to other prediction tasks and learning objectives.
| stat |
Spatial Heterogeneity Automatic Detection and Estimation | Spatial regression is widely used for modeling the relationship between a
dependent variable and explanatory covariates. Oftentimes, the linear
relationships vary across space, when some covariates have location-specific
effects on the response. One fundamental question is how to detect the
systematic variation in the model and identify which locations share common
regression coefficients and which do not. Only a correct model structure can
assure unbiased estimation of coefficients and valid inferences. In this work,
we propose a new procedure, called Spatial Heterogeneity Automatic Detection
and Estimation (SHADE), for automatically and simultaneously subgrouping and
estimating covariate effects for spatial regression models. The SHADE employs a
class of spatially-weighted fusion type penalty on all pairs of observations,
with location-specific weight adaptively constructed using spatial information,
to cluster coefficients into subgroups. Under certain regularity conditions,
the SHADE is shown to be able to identify the true model structure with
probability approaching one and estimate regression coefficients consistently.
We develop an alternating direction method of multiplier algorithm (ADMM) to
compute the SHAD efficiently. In numerical studies, we demonstrate empirical
performance of the SHADE by using different choices of weights and compare
their accuracy. The results suggest that spatial information can enhance
subgroup structure analysis in challenging situations when the spatial
variation among regression coefficients is small or the number of repeated
measures is small. Finally, the SHADE is applied to find the relationship
between a natural resource survey and a land cover data layer to identify
spatially interpretable groups.
| stat |
An end-to-end data-driven optimisation framework for constrained
trajectories | Many real-world problems require to optimise trajectories under constraints.
Classical approaches are based on optimal control methods but require an exact
knowledge of the underlying dynamics, which could be challenging or even out of
reach. In this paper, we leverage data-driven approaches to design a new
end-to-end framework which is dynamics-free for optimised and realistic
trajectories. We first decompose the trajectories on function basis, trading
the initial infinite dimension problem on a multivariate functional space for a
parameter optimisation problem. A maximum \emph{a posteriori} approach which
incorporates information from data is used to obtain a new optimisation problem
which is regularised. The penalised term focuses the search on a region
centered on data and includes estimated linear constraints in the problem. We
apply our data-driven approach to two settings in aeronautics and sailing
routes optimisation, yielding commanding results. The developed approach has
been implemented in the Python library PyRotor.
| stat |
A Theory of Dichotomous Valuation with Applications to Variable
Selection | An econometric or statistical model may undergo a marginal gain if we admit a
new variable to the model, and a marginal loss if we remove an existing
variable from the model. Assuming equality of opportunity among all candidate
variables, we derive a valuation framework by the expected marginal gain and
marginal loss in all potential modeling scenarios. However, marginal gain and
loss are not symmetric; thus, we introduce three unbiased solutions. When used
in variable selection, our new approaches significantly outperform several
popular methods used in practice. The results also explore some novel traits of
the Shapley value.
| stat |
Data-Based Optimal Bandwidth for Kernel Density Estimation of
Statistical Samples | It is a common practice to evaluate probability density function or matter
spatial density function from statistical samples. Kernel density estimation is
a frequently used method, but to select an optimal bandwidth of kernel
estimation, which is completely based on data samples, is a long-term issue
that has not been well settled so far. There exist analytic formulae of optimal
kernel bandwidth, but they cannot be applied directly to data samples, since
they depend on the unknown underlying density functions from which the samples
are drawn. In this work, we devise an approach to pick out the totally
data-based optimal bandwidth. First, we derive correction formulae for the
analytic formulae of optimal bandwidth to compute the roughness of the sample's
density function. Then substitute the correction formulae into the analytic
formulae for optimal bandwidth, and through iteration, we obtain the sample's
optimal bandwidth. Compared with analytic formulae, our approach gives very
good results, with relative differences from the analytic formulae being only
2%-3% for a sample size larger than 10^4. This approach can also be generalized
easily to cases of variable kernel estimations.
| stat |
Distributed ARIMA Models for Ultra-long Time Series | Providing forecasts for ultra-long time series plays a vital role in various
activities, such as investment decisions, industrial production arrangements,
and farm management. This paper develops a novel distributed forecasting
framework to tackle challenges associated with forecasting ultra-long time
series by utilizing the industry-standard MapReduce framework. The proposed
model combination approach facilitates distributed time series forecasting by
combining the local estimators of ARIMA (AutoRegressive Integrated Moving
Average) models delivered from worker nodes and minimizing a global loss
function. In this way, instead of unrealistically assuming the data generating
process (DGP) of an ultra-long time series stays invariant, we make assumptions
only on the DGP of subseries spanning shorter time periods. We investigate the
performance of the proposed distributed ARIMA models on an electricity demand
dataset. Compared to ARIMA models, our approach results in significantly
improved forecasting accuracy and computational efficiency both in point
forecasts and prediction intervals, especially for longer forecast horizons.
Moreover, we explore some potential factors that may affect the forecasting
performance of our approach.
| stat |
Combining Population and Study Data for Inference on Event Rates | This note considers the problem of conducting statistical inference on the
share of individuals in some subgroup of a population that experience some
event. The specific complication is that the size of the subgroup needs to be
estimated, whereas the number of individuals that experience the event is
known. The problem is motivated by the recent study of Streeck et al. (2020),
who estimate the infection fatality rate (IFR) of SARS-CoV-2 infection in a
German town that experienced a super-spreading event in mid-February 2020. In
their case the subgroup of interest is comprised of all infected individuals,
and the event is death caused by the infection. We clarify issues with the
precise definition of the target parameter in this context, and propose
confidence intervals (CIs) based on classical statistical principles that
result in good coverage properties.
| stat |
Convergence diagnostics for Markov chain Monte Carlo | Markov chain Monte Carlo (MCMC) is one of the most useful approaches to
scientific computing because of its flexible construction, ease of use and
generality. Indeed, MCMC is indispensable for performing Bayesian analysis. Two
critical questions that MCMC practitioners need to address are where to start
and when to stop the simulation. Although a great amount of research has gone
into establishing convergence criteria and stopping rules with sound
theoretical foundation, in practice, MCMC users often decide convergence by
applying empirical diagnostic tools. This review article discusses the most
widely used MCMC convergence diagnostic tools. Some recently proposed stopping
rules with firm theoretical footing are also presented. The convergence
diagnostics and stopping rules are illustrated using three detailed examples.
| stat |
On the bias of H-scores for comparing biclusters, and how to correct it | In the last two decades several biclustering methods have been developed as
new unsupervised learning techniques to simultaneously cluster rows and columns
of a data matrix. These algorithms play a central role in contemporary machine
learning and in many applications, e.g. to computational biology and
bioinformatics. The H-score is the evaluation score underlying the seminal
biclustering algorithm by Cheng and Church, as well as many other subsequent
biclustering methods. In this paper, we characterize a potentially troublesome
bias in this score, that can distort biclustering results. We prove, both
analytically and by simulation, that the average H-score increases with the
number of rows/columns in a bicluster. This makes the H-score, and hence all
algorithms based on it, biased towards small clusters. Based on our analytical
proof, we are able to provide a straightforward way to correct this bias,
allowing users to accurately compare biclusters.
| stat |
Randomization for the susceptibility effect of an infectious disease
intervention | Randomized trials of infectious disease interventions, such as vaccines,
often focus on groups of connected or potentially interacting individuals. When
the pathogen of interest is transmissible between study subjects, interference
may occur: individual infection outcomes may depend on treatments received by
others. Epidemiologists have defined the primary causal effect of interest --
called the "susceptibility effect" -- as a contrast in infection risk under
treatment versus no treatment, while holding exposure to infectiousness
constant. A related quantity -- the "direct effect" -- is defined as an
unconditional contrast between the infection risk under treatment versus no
treatment. The purpose of this paper is to show that under a widely recommended
randomization design, the direct effect may fail to recover the sign of the
true susceptibility effect of the intervention in a randomized trial when
outcomes are contagious. The analytical approach uses structural features of
infectious disease transmission to define the susceptibility effect. A new
probabilistic coupling argument reveals stochastic dominance relations between
potential infection outcomes under different treatment allocations. The results
suggest that estimating the direct effect under randomization may provide
misleading inferences about the effect of an intervention -- such as a vaccine
-- when outcomes are contagious.
| stat |
A general theory of regression adjustment for covariate-adaptive
randomization: OLS, Lasso, and beyond | We consider the problem of estimating and inferring treatment effects in
randomized experiments. In practice, stratified randomization, or more
generally, covariate-adaptive randomization, is routinely used in the design
stage to balance the treatment allocations with respect to a few variables that
are most relevant to the outcomes. Then, regression is performed in the
analysis stage to adjust the remaining imbalances to yield more efficient
treatment effect estimators. Building upon and unifying the recent results
obtained for ordinary least squares adjusted estimators under
covariate-adaptive randomization, this paper presents a general theory of
regression adjustment that allows for arbitrary model misspecification and the
presence of a large number of baseline covariates. We exemplify the theory on
two Lasso-adjusted treatment effect estimators, both of which are optimal in
their respective classes. In addition, nonparametric consistent variance
estimators are proposed to facilitate valid inferences, which work irrespective
of the specific randomization methods used. The robustness and improved
efficiency of the proposed estimators are demonstrated through a simulation
study and a clinical trial example. This study sheds light on improving
treatment effect estimation efficiency by implementing machine learning methods
in covariate-adaptive randomized experiments.
| stat |
Mixed effects models for healthcare longitudinal data with an
informative visiting process: a Monte Carlo simulation study | Electronic health records are being increasingly used in medical research to
answer more relevant and detailed clinical questions; however, they pose new
and significant methodological challenges. For instance, observation times are
likely correlated with the underlying disease severity: patients with worse
conditions utilise health care more and may have worse biomarker values
recorded. Traditional methods for analysing longitudinal data assume
independence between observation times and disease severity; yet, with
healthcare data such assumptions unlikely holds. Through Monte Carlo
simulation, we compare different analytical approaches proposed to account for
an informative visiting process to assess whether they lead to unbiased
results. Furthermore, we formalise a joint model for the observation process
and the longitudinal outcome within an extended joint modelling framework. We
illustrate our results using data from a pragmatic trial on enhanced care for
individuals with chronic kidney disease, and we introduce user-friendly
software that can be used to fit the joint model for the observation process
and a longitudinal outcome.
| stat |
On the Adversarial Robustness of Multivariate Robust Estimation | In this paper, we investigate the adversarial robustness of multivariate
$M$-Estimators. In the considered model, after observing the whole dataset, an
adversary can modify all data points with the goal of maximizing inference
errors. We use adversarial influence function (AIF) to measure the asymptotic
rate at which the adversary can change the inference result. We first
characterize the adversary's optimal modification strategy and its
corresponding AIF. From the defender's perspective, we would like to design an
estimator that has a small AIF. For the case of joint location and scale
estimation problem, we characterize the optimal $M$-estimator that has the
smallest AIF. We further identify a tradeoff between robustness against
adversarial modifications and robustness against outliers, and derive the
optimal $M$-estimator that achieves the best tradeoff.
| stat |
Nonparametric Multivariate Density Estimation: A Low-Rank Characteristic
Function Approach | Effective non-parametric density estimation is a key challenge in
high-dimensional multivariate data analysis. In this paper,we propose a novel
approach that builds upon tensor factorization tools. Any multivariate density
can be represented by its characteristic function, via the Fourier transform.
If the sought density is compactly supported, then its characteristic function
can be approximated, within controllable error, by a finite tensor of leading
Fourier coefficients, whose size de-pends on the smoothness of the underlying
density. This tensor can be naturally estimated from observed realizations of
the random vector of interest, via sample averaging. In order to circumvent the
curse of dimensionality, we introduce a low-rank model of this characteristic
tensor, which significantly improves the density estimate especially for
high-dimensional data and/or in the sample-starved regime. By virtue of
uniqueness of low-rank tensor decomposition, under certain conditions, our
method enables learning the true data-generating distribution. We demonstrate
the very promising performance of the proposed method using several measured
datasets.
| stat |
Scaling Hamiltonian Monte Carlo Inference for Bayesian Neural Networks
with Symmetric Splitting | Hamiltonian Monte Carlo (HMC) is a Markov chain Monte Carlo (MCMC) approach
that exhibits favourable exploration properties in high-dimensional models such
as neural networks. Unfortunately, HMC has limited use in large-data regimes
and little work has explored suitable approaches that aim to preserve the
entire Hamiltonian. In our work, we introduce a new symmetric integration
scheme for split HMC that does not rely on stochastic gradients. We show that
our new formulation is more efficient than previous approaches and is easy to
implement with a single GPU. As a result, we are able to perform full HMC over
common deep learning architectures using entire data sets. In addition, when we
compare with stochastic gradient MCMC, we show that our method achieves better
performance in both accuracy and uncertainty quantification. Our approach
demonstrates HMC as a feasible option when considering inference schemes for
large-scale machine learning problems.
| stat |
A simple recipe for making accurate parametric inference in finite
sample | Constructing tests or confidence regions that control over the error rates in
the long-run is probably one of the most important problem in statistics. Yet,
the theoretical justification for most methods in statistics is asymptotic. The
bootstrap for example, despite its simplicity and its widespread usage, is an
asymptotic method. There are in general no claim about the exactness of
inferential procedures in finite sample. In this paper, we propose an
alternative to the parametric bootstrap. We setup general conditions to
demonstrate theoretically that accurate inference can be claimed in finite
sample.
| stat |
Scalable Low-Rank Autoregressive Tensor Learning for Spatiotemporal
Traffic Data Imputation | Missing value problem in spatiotemporal traffic data has long been a
challenging topic, in particular for large-scale and high-dimensional data with
complex missing mechanisms and diverse degrees of missingness. Recent studies
based on tensor nuclear norm have demonstrated the superiority of tensor
learning in imputation tasks by effectively characterizing the complex
correlations/dependencies in spatiotemporal data. However, despite the
promising results, these approaches do not scale well to large tensors. In this
paper, we focus on addressing the missing data imputation problem for
large-scale spatiotemporal traffic data. To achieve both high accuracy and
efficiency, we develop a scalable autoregressive tensor learning
model---Low-Tubal-Rank Autoregressive Tensor Completion (LATC-Tubal)---based on
the existing framework of Low-Rank Autoregressive Tensor Completion (LATC),
which is well-suited for spatiotemporal traffic data that characterized by
multidimensional structure of location$\times$ time of day $\times$ day. In
particular, the proposed LATC-Tubal model involves a scalable tensor nuclear
norm minimization scheme by integrating linear unitary transformation.
Therefore, the tensor nuclear norm minimization can be solved by singular value
thresholding on the transformed matrix of each day while the day-to-day
correlation can be effectively preserved by the unitary transform matrix.
Before setting up the experiment, we consider two large-scale 5-minute traffic
speed data sets collected by the California PeMS system with 11160 sensors. We
compare LATC-Tubal with state-of-the-art baseline models, and find that
LATC-Tubal can achieve competitively accuracy with a significantly lower
computational cost. In addition, the LATC-Tubal will also benefit other tasks
in modeling large-scale spatiotemporal traffic data, such as network-level
traffic forecasting.
| stat |
Topological Data Analysis of Database Representations for Information
Retrieval | Appropriately representing elements in a database so that queries may be
accurately matched is a central task in information retrieval. This recently
has been achieved by embedding the graphical structure of the database into a
manifold so that the hierarchy is preserved. Persistent homology provides a
rigorous characterization for the database topology in terms of both its
hierarchy and connectivity structure. We compute persistent homology on a
variety of datasets and show that some commonly used embeddings fail to
preserve the connectivity. Moreover, we show that embeddings which successfully
retain the database topology coincide in persistent homology. We introduce the
dilation-invariant bottleneck distance to capture this effect, which addresses
metric distortion on manifolds. We use it to show that distances between
topology-preserving embeddings of databases are small.
| stat |
Log-symmetric quantile regression models | Regression models based on the log-symmetric family of distributions are
particularly useful when the response is strictly positive and asymmetric. In
this paper, we propose a class of quantile regression models based on
reparameterized log-symmetric distributions, which have a quantile parameter.
Two Monte Carlo simulation studies are carried out using the R software. The
first one analyzes the performance of the maximum likelihood estimators, the
information criteria AIC, BIC and AICc, and the generalized Cox-Snell and
random quantile residuals. The second one evaluates the performance of the size
and power of the Wald, likelihood ratio, score and gradient tests. A real box
office data set is finally analyzed to illustrate the proposed approach.
| stat |
Ultrahigh-dimensional Robust and Efficient Sparse Regression using
Non-Concave Penalized Density Power Divergence | We propose a sparse regression method based on the non-concave penalized
density power divergence loss function which is robust against infinitesimal
contamination in very high dimensionality. Present methods of sparse and robust
regression are based on $\ell_1$-penalization, and their theoretical properties
are not well-investigated. In contrast, we use a general class of folded
concave penalties that ensure sparse recovery and consistent estimation of
regression coefficients. We propose an alternating algorithm based on the
Concave-Convex procedure to obtain our estimate, and demonstrate its robustness
properties using influence function analysis. Under some conditions on the
fixed design matrix and penalty function, we prove that this estimator
possesses large-sample oracle properties in an ultrahigh-dimensional regime.
The performance and effectiveness of our proposed method for parameter
estimation and prediction compared to state-of-the-art are demonstrated through
simulation studies.
| stat |
No-harm calibration for generalized Oaxaca-Blinder estimators | In randomized experiments, linear regression with baseline features can be
used to form an estimate of the sample average treatment effect that is
asymptotically no less efficient than the treated-minus-control difference in
means. Randomization alone provides this "do-no-harm" property, with neither
truth of a linear model nor a generative model for the outcomes being required.
We present a general calibration step which confers the same no-harm property
onto estimators leveraging a broad class of nonlinear models. The process
recovers the usual regression-adjusted estimator when ordinary least squares is
used, and further provides non-inferior treatment effect estimators using
methods such as logistic and Poisson regression. The resulting estimators are
non-inferior with respect to both the difference in means estimator and with
respect to treatment effect estimators that have not undergone calibration.
| stat |
Bayesian design and analysis of external pilot trials for complex
interventions | External pilot trials of complex interventions are used to help determine if
and how a confirmatory trial should be undertaken, providing estimates of
parameters such as recruitment, retention and adherence rates. The decision to
progress to the confirmatory trial is typically made by comparing these
estimates to pre-specified thresholds known as progression criteria, although
the statistical properties of such decision rules are rarely assessed. Such
assessment is complicated by several methodological challenges, including the
simultaneous evaluation of multiple endpoints, complex multi-level models,
small sample sizes, and uncertainty in nuisance parameters. In response to
these challenges, we describe a Bayesian approach to the design and analysis of
external pilot trials. We show how progression decisions can be made by
minimising the expected value of a loss function, defined over the whole
parameter space to allow for preferences and trade-offs between multiple
parameters to be articulated and used in the decision making process. The
assessment of preferences is kept feasible by using a piecewise constant
parameterisation of the loss function, the parameters of which are chosen at
the design stage to lead to desirable operating characteristics. We describe a
flexible, yet computationally intensive, nested Monte Carlo algorithm for
estimating operating characteristics. The method is used to revisit the design
of an external pilot trial of a complex intervention designed to increase the
physical activity of care home residents.
| stat |
Computation of Standardized Residuals for MARSS Models | This report shows how to compute the variance of the joint conditional model
and state residuals for multivariate autoregressive Gaussian state-space
(MARSS) models. The MARSS model can be written: x(t)=Bx(t-1)+u+w(t),
y(t)=Zx(t)+a+v(t), where w(t) and v(t) are multivariate normal error-terms with
variance-covariance matrices Q and R respectively. The joint conditional
residuals are the w(t) and v(t) conditioned on a set of, possibly incomplete,
data y. Harvey, Koopman and Penzer (1998) show a recursive algorithm for these
residuals. I show the equation for the residuals using the conditional
variances of the states and the conditional covariance between unobserved data
and states. This allows one to compute the variance of un-observed residuals,
which could be useful for leave-one-out cross-validation tests. I also show how
to modify the Harvey et al algorithm in the case of missing values and how to
modify it to return the non-normalized conditional residuals.
| stat |
Optimal Uncertainty Quantification of a risk measurement from a
thermal-hydraulic code using Canonical Moments | We study an industrial computer code related to nuclear safety. A major topic
of interest is to assess the uncertainties tainting the results of a computer
simulation. In this work we gain robustness on the quantification of a risk
measurement by accounting for all sources of uncertainties tainting the inputs
of a computer code. To that extent, we evaluate the maximum quantile over a
class of distributions defined only by constraints on their moments. Two
options are available when dealing with such complex optimization problems: one
can either optimize under constraints; or preferably, one should reformulate
the objective function. We identify a well suited parameterization to compute
the optimal quantile based on the theory of canonical moments. It allows an
effective, free of constraints, optimization.
| stat |
BIMC: The Bayesian Inverse Monte Carlo method for goal-oriented
uncertainty quantification. Part II | In Part I (arXiv:1911.00619) of this article, we proposed an importance
sampling algorithm to compute rare-event probabilities in forward uncertainty
quantification problems. The algorithm, which we termed the "Bayesian Inverse
Monte Carlo (BIMC) method", was shown to be optimal for problems in which the
input-output operator is nearly linear. But applying the original BIMC to
highly nonlinear systems can lead to several different failure modes. In this
paper, we modify the BIMC method to extend its applicability to a wider class
of systems. The modified algorithm, which we call "Adaptive-BIMC (A-BIMC)", has
two stages. In the first stage, we solve a sequence of optimization problems to
roughly identify those regions of parameter space which trigger the rare-event.
In the second stage, we use the stage one results to construct a mixture of
Gaussians that can be then used in an importance sampling algorithm to estimate
rare event probability. We propose using a local surrogate that minimizes
costly forward solves. The effectiveness of A-BIMC is demonstrated via several
synthetic examples. Yet again, the modified algorithm is prone to failure. We
systematically identify conditions under which it fails to lead to an effective
importance sampling distribution.
| stat |
Noise and Fluctuation of Finite Learning Rate Stochastic Gradient
Descent | In the vanishing learning rate regime, stochastic gradient descent (SGD) is
now relatively well understood. In this work, we propose to study the basic
properties of SGD and its variants in the non-vanishing learning rate regime.
The focus is on deriving exactly solvable results and discussing their
implications. The main contributions of this work are to derive the stationary
distribution for discrete-time SGD in a quadratic loss function with and
without momentum; in particular, one implication of our result is that the
fluctuation caused by discrete-time dynamics takes a distorted shape and is
dramatically larger than a continuous-time theory could predict. Examples of
applications of the proposed theory considered in this work include the
approximation error of variants of SGD, the effect of minibatch noise, the
optimal Bayesian inference, the escape rate from a sharp minimum, and the
stationary distribution of a few second-order methods including damped Newton's
method and natural gradient descent.
| stat |
Exploring data subsets with vtree | Variable trees are a new method for the exploration of discrete multivariate
data. They display nested subsets and corresponding frequencies and
percentages. Manual calculation of these quantities can be laborious,
especially when there are many multi-level factors and missing data. Here we
introduce variable trees and their implementation in the vtree R package, draw
comparisons with existing methods (contingency tables, mosaic plots, Venn/Euler
diagrams, and UpSet), and illustrate their utility using two case studies.
Variable trees can be used to (1) reveal patterns in nested subsets, (2)
explore missing data, and (3) generate study flow diagrams (e.g., CONSORT
diagrams) directly from data frames, to support reproducible research and open
science.
| stat |
ACVAE-VC: Non-parallel many-to-many voice conversion with auxiliary
classifier variational autoencoder | This paper proposes a non-parallel many-to-many voice conversion (VC) method
using a variant of the conditional variational autoencoder (VAE) called an
auxiliary classifier VAE (ACVAE). The proposed method has three key features.
First, it adopts fully convolutional architectures to construct the encoder and
decoder networks so that the networks can learn conversion rules that capture
time dependencies in the acoustic feature sequences of source and target
speech. Second, it uses an information-theoretic regularization for the model
training to ensure that the information in the attribute class label will not
be lost in the conversion process. With regular CVAEs, the encoder and decoder
are free to ignore the attribute class label input. This can be problematic
since in such a situation, the attribute class label will have little effect on
controlling the voice characteristics of input speech at test time. Such
situations can be avoided by introducing an auxiliary classifier and training
the encoder and decoder so that the attribute classes of the decoder outputs
are correctly predicted by the classifier. Third, it avoids producing
buzzy-sounding speech at test time by simply transplanting the spectral details
of the input speech into its converted version. Subjective evaluation
experiments revealed that this simple method worked reasonably well in a
non-parallel many-to-many speaker identity conversion task.
| stat |
Linear screening for high-dimensional computer experiments | In this paper we propose a linear variable screening method for computer
experiments when the number of input variables is larger than the number of
runs. This method uses a linear model to model the nonlinear data, and screens
the important variables by existing screening methods for linear models. When
the underlying simulator is nearly sparse, we prove that the linear screening
method is asymptotically valid under mild conditions. To improve the screening
accuracy, we also provide a two-stage procedure that uses different basis
functions in the linear model. The proposed methods are very simple and easy to
implement. Numerical results indicate that our methods outperform existing
model-free screening methods.
| stat |
Scalable visualisation methods for modern Generalized Additive Models | In the last two decades the growth of computational resources has made it
possible to handle Generalized Additive Models (GAMs) that formerly were too
costly for serious applications. However, the growth in model complexity has
not been matched by improved visualisations for model development and results
presentation. Motivated by an industrial application in electricity load
forecasting, we identify the areas where the lack of modern visualisation tools
for GAMs is particularly severe, and we address the shortcomings of existing
methods by proposing a set of visual tools that a) are fast enough for
interactive use, b) exploit the additive structure of GAMs, c) scale to large
data sets and d) can be used in conjunction with a wide range of response
distributions. All the new visual methods proposed in this work are implemented
by the mgcViz R package, which can be found on the Comprehensive R Archive
Network.
| stat |
Nonparametric Functional Approximation with Delaunay Triangulation | We propose a differentiable nonparametric algorithm, the Delaunay
triangulation learner (DTL), to solve the functional approximation problem on
the basis of a $p$-dimensional feature space. By conducting the Delaunay
triangulation algorithm on the data points, the DTL partitions the feature
space into a series of $p$-dimensional simplices in a geometrically optimal
way, and fits a linear model within each simplex. We study its theoretical
properties by exploring the geometric properties of the Delaunay triangulation,
and compare its performance with other statistical learners in numerical
studies.
| stat |
Towards More Flexible False Positive Control in Phase III Randomized
Clinical Trials | Phase III randomized clinical trials play a monumentally critical role in the
evaluation of new medical products. Because of the intrinsic nature of
uncertainty embedded in our capability in assessing the efficacy of a medical
product, interpretation of trial results relies on statistical principles to
control the error of false positives below desirable level. The
well-established statistical hypothesis testing procedure suffers from two
major limitations, namely, the lack of flexibility in the thresholds to claim
success and the lack of capability of controlling the total number of false
positives that could be yielded by the large volume of trials. We propose two
general theoretical frameworks based on the conventional frequentist paradigm
and Bayesian perspectives, which offer realistic, flexible and effective
solutions to these limitations. Our methods are based on the distribution of
the effect sizes of the population of trials of interest. The estimation of
this distribution is practically feasible as clinicaltrials.gov provides a
centralized data repository with unbiased coverage of clinical trials. We
provide a detailed development of the two frameworks with numerical results
obtained for industry sponsored Phase III randomized clinical trials.
| stat |
On the Convergence of Stochastic Gradient Descent with Adaptive
Stepsizes | Stochastic gradient descent is the method of choice for large scale
optimization of machine learning objective functions. Yet, its performance is
greatly variable and heavily depends on the choice of the stepsizes. This has
motivated a large body of research on adaptive stepsizes. However, there is
currently a gap in our theoretical understanding of these methods, especially
in the non-convex setting. In this paper, we start closing this gap: we
theoretically analyze in the convex and non-convex settings a generalized
version of the AdaGrad stepsizes. We show sufficient conditions for these
stepsizes to achieve almost sure asymptotic convergence of the gradients to
zero, proving the first guarantee for generalized AdaGrad stepsizes in the
non-convex setting. Moreover, we show that these stepsizes allow to
automatically adapt to the level of noise of the stochastic gradients in both
the convex and non-convex settings, interpolating between $O(1/T)$ and
$O(1/\sqrt{T})$, up to logarithmic terms.
| stat |
Dynamic Silos: Modularity in intra-organizational communication networks
during the Covid-19 pandemic | Workplace communications around the world were drastically altered by
Covid-19, work-from-home orders, and the rise of remote work. We analyze
aggregated, anonymized metadata from over 360 billion emails within over 4000
organizations worldwide to examine changes in network community structures from
2019 through 2020. We find that, during 2020, organizations around the world
became more siloed, evidenced by increased modularity. This shift was
concurrent with decreased stability, indicating that organizational siloes had
less stable membership. We provide initial insights into the implications of
these network changes -- which we term dynamic silos -- for organizational
performance and innovation.
| stat |
Multilevel Dimension-Independent Likelihood-Informed MCMC for
Large-Scale Inverse Problems | We present a non-trivial integration of dimension-independent
likelihood-informed (DILI) MCMC (Cui, Law, Marzouk, 2016) and the multilevel
MCMC (Dodwell et al., 2015) to explore the hierarchy of posterior
distributions. This integration offers several advantages: First, DILI-MCMC
employs an intrinsic likelihood-informed subspace (LIS) (Cui et al., 2014) --
which involves a number of forward and adjoint model simulations -- to design
accelerated operator-weighted proposals. By exploiting the multilevel structure
of the discretised parameters and discretised forward models, we design a
Rayleigh-Ritz procedure to significantly reduce the computational effort in
building the LIS and operating with DILI proposals. Second, the resulting
DILI-MCMC can drastically improve the sampling efficiency of MCMC at each
level, and hence reduce the integration error of the multilevel algorithm for
fixed CPU time. To be able to fully exploit the power of multilevel MCMC and to
reduce the dependencies of samples on different levels for a parallel
implementation, we also suggest a new pooling strategy for allocating
computational resources across different levels and constructing Markov chains
at higher levels conditioned on those simulated on lower levels. Numerical
results confirm the improved computational efficiency of the multilevel DILI
approach.
| stat |
Inferring the unknown parameters in Differential Equation by Gaussian
Process Regression with Constraint | Differential Equation (DE) is a commonly used modeling method in various
scientific subjects such as finance and biology. The parameters in DE models
often have interesting scientific interpretations, but their values are often
unknown and need to be estimated from the measurements of the DE. In this work,
we propose a Bayesian inference framework to solve the problem of estimating
the parameters of the DE model, from the given noisy and scarce observations of
the solution only. A key issue in this problem is to robustly estimate the
derivatives of a function from noisy observations of only the function values
at given location points, under the assumption of a physical model in the form
of differential equation governing the function and its derivatives. To address
the key issue, we use the Gaussian Process Regression with Constraint (GPRC)
method which jointly model the solution, the derivatives, and the parametric
differential equation, to estimate the solution and its derivatives. For
nonlinear differential equations, a Picard-iteration-like approximation of
linearization method is used so that the GPRC can be still iteratively
applicable. A new potential which combines the data and equation information,
is proposed and used in the likelihood for our inference. With numerical
examples, we illustrate that the proposed method has competitive performance
against existing approaches for estimating the unknown parameters in DEs.
| stat |
Survival analysis as a classification problem | In this paper, we explore a method for treating survival analysis as a
classification problem. The method uses a "stacking" idea that collects the
features and outcomes of the survival data in a large data frame, and then
treats it as a classification problem. In this framework, various statistical
learning algorithms (including logistic regression, random forests, gradient
boosting machines and neural networks) can be applied to estimate the
parameters and make predictions. For stacking with logistic regression, we show
that this approach is approximately equivalent to the Cox proportional hazards
model with both theoretical analysis and simulation studies. For stacking with
other machine learning algorithms, we show through simulation studies that our
method can outperform Cox proportional hazards model in terms of estimated
survival curves. This idea is not new, but we believe that it should be better
known by statistiicians and other data scientists.
| stat |
Net benefit separation and the determination curve: a probabilistic
framework for cost-effectiveness estimation | Considerations regarding clinical effectiveness and cost are essential in
comparing the overall value of two treatments. There has been growing interest
in methodology to integrate cost and effectiveness measures in order to inform
policy and promote adequate resource allocation. The net monetary benefit
aggregates information on differences in mean cost and clinical outcomes; the
cost-effectiveness acceptability curve was then developed to characterize the
extent to which the strength of evidence regarding net monetary benefit changes
with fluctuations in the willingness-to-pay threshold. Methods to derive
insights from characteristics of the cost/clinical outcomes besides mean
differences remain undeveloped but may also be informative. We propose a novel
probabilistic measure of cost-effectiveness based on the stochastic ordering of
the individual net benefit distribution under each treatment. Our approach is
able to accommodate features frequently encountered in observational data
including confounding and censoring, and complements the net monetary benefit
in the insights it provides. We conduct a range of simulations to evaluate
finite-sample performance and illustrate our proposed approach using simulated
data based on a study of endometrial cancer patients.
| stat |
On the Heavy-Tailed Theory of Stochastic Gradient Descent for Deep
Neural Networks | The gradient noise (GN) in the stochastic gradient descent (SGD) algorithm is
often considered to be Gaussian in the large data regime by assuming that the
\emph{classical} central limit theorem (CLT) kicks in. This assumption is often
made for mathematical convenience, since it enables SGD to be analyzed as a
stochastic differential equation (SDE) driven by a Brownian motion. We argue
that the Gaussianity assumption might fail to hold in deep learning settings
and hence render the Brownian motion-based analyses inappropriate. Inspired by
non-Gaussian natural phenomena, we consider the GN in a more general context
and invoke the \emph{generalized} CLT, which suggests that the GN converges to
a \emph{heavy-tailed} $\alpha$-stable random vector, where \emph{tail-index}
$\alpha$ determines the heavy-tailedness of the distribution. Accordingly, we
propose to analyze SGD as a discretization of an SDE driven by a L\'{e}vy
motion. Such SDEs can incur `jumps', which force the SDE and its discretization
\emph{transition} from narrow minima to wider minima, as proven by existing
metastability theory and the extensions that we proved recently. In this study,
under the $\alpha$-stable GN assumption, we further establish an explicit
connection between the convergence rate of SGD to a local minimum and the
tail-index $\alpha$. To validate the $\alpha$-stable assumption, we conduct
experiments on common deep learning scenarios and show that in all settings,
the GN is highly non-Gaussian and admits heavy-tails. We investigate the tail
behavior in varying network architectures and sizes, loss functions, and
datasets. Our results open up a different perspective and shed more light on
the belief that SGD prefers wide minima.
| stat |
Identification of Causal Diffusion Effects Under Structural Stationarity | Although social and biomedical scientists have long been interested in the
process through which ideas and behaviors diffuse, the identification of causal
diffusion effects, also known as peer and contagion effects, remains
challenging. Many scholars consider the commonly used assumption of no omitted
confounders to be untenable due to contextual confounding and homophily bias.
To address this long-standing problem, we examine the causal identification
under a new assumption of structural stationarity, which formalizes the
underlying diffusion process with a class of dynamic causal directed acyclic
graphs. First, we develop a statistical test that can detect a wide range of
biases, including the two types mentioned above. We then propose a
difference-in-differences style estimator that can directly correct biases
under an additional parametric assumption. Leveraging the proposed methods, we
study the spatial diffusion of hate crimes against refugees in Germany. After
correcting large upward bias in existing studies, we find hate crimes diffuse
only to areas that have a high proportion of school dropouts.
| stat |
Integrated Time Series Summarization and Prediction Algorithm and its
Application to COVID-19 Data Mining | This paper proposes a simple method to extract from a set of multiple related
time series a compressed representation for each time series based on
statistics for the entire set of all time series. This is achieved by a
hierarchical algorithm that first generates an alphabet of shapelets based on
the segmentation of centroids for clustered data, before labels of these
shapelets are assigned to the segmentation of each single time series via
nearest neighbor search using unconstrained dynamic time warping as distance
measure to deal with non-uniform time series lenghts. Thereby, a sequence of
labels is assigned for each time series. Completion of the last label sequence
permits prediction of individual time series. Proposed method is evaluated on
two global COVID-19 datasets, first, for the number of daily net cases (daily
new infections minus daily recoveries), and, second, for the number of daily
deaths attributed to COVID-19 as of April 27, 2020. The first dataset involves
249 time series for different countries, each of length 96. The second dataset
involves 264 time series, each of length 96. Based on detected anomalies in
available data a decentralized exit strategy from lockdowns is advocated.
| stat |
Estimating Full Lipschitz Constants of Deep Neural Networks | We estimate the Lipschitz constants of the gradient of a deep neural network
and the network itself with respect to the full set of parameters. We first
develop estimates for a deep feed-forward densely connected network and then,
in a more general framework, for all neural networks that can be represented as
solutions of controlled ordinary differential equations, where time appears as
continuous depth. These estimates can be used to set the step size of
stochastic gradient descent methods, which is illustrated for one example
method.
| stat |
Accelerating delayed-acceptance Markov chain Monte Carlo algorithms | Delayed-acceptance Markov chain Monte Carlo (DA-MCMC) samples from a
probability distribution via a two-stages version of the Metropolis-Hastings
algorithm, by combining the target distribution with a "surrogate" (i.e. an
approximate and computationally cheaper version) of said distribution. DA-MCMC
accelerates MCMC sampling in complex applications, while still targeting the
exact distribution. We design a computationally faster, albeit approximate,
DA-MCMC algorithm. We consider parameter inference in a Bayesian setting where
a surrogate likelihood function is introduced in the delayed-acceptance scheme.
When the evaluation of the likelihood function is computationally intensive,
our scheme produces a 2-4 times speed-up, compared to standard DA-MCMC.
However, the acceleration is highly problem dependent. Inference results for
the standard delayed-acceptance algorithm and our approximated version are
similar, indicating that our algorithm can return reliable Bayesian inference.
As a computationally intensive case study, we introduce a novel stochastic
differential equation model for protein folding data.
| stat |
Pseudo-Bayesian Learning with Kernel Fourier Transform as Prior | We revisit Rahimi and Recht (2007)'s kernel random Fourier features (RFF)
method through the lens of the PAC-Bayesian theory. While the primary goal of
RFF is to approximate a kernel, we look at the Fourier transform as a prior
distribution over trigonometric hypotheses. It naturally suggests learning a
posterior on these hypotheses. We derive generalization bounds that are
optimized by learning a pseudo-posterior obtained from a closed-form
expression. Based on this study, we consider two learning strategies: The first
one finds a compact landmarks-based representation of the data where each
landmark is given by a distribution-tailored similarity measure, while the
second one provides a PAC-Bayesian justification to the kernel alignment method
of Sinha and Duchi (2016).
| stat |
Reducing Sampling Ratios Improves Bagging in Sparse Regression | Bagging, a powerful ensemble method from machine learning, improves the
performance of unstable predictors. Although the power of Bagging has been
shown mostly in classification problems, we demonstrate the success of
employing Bagging in sparse regression over the baseline method (L1
minimization). The framework employs the generalized version of the original
Bagging with various bootstrap ratios. The performance limits associated with
different choices of bootstrap sampling ratio L/m and number of estimates K is
analyzed theoretically. Simulation shows that the proposed method yields
state-of-the-art recovery performance, outperforming L1 minimization and
Bolasso in the challenging case of low levels of measurements. A lower L/m
ratio (60% - 90%) leads to better performance, especially with a small number
of measurements. With the reduced sampling rate, SNR improves over the original
Bagging by up to 24%. With a properly chosen sampling ratio, a reasonably small
number of estimates K = 30 gives satisfying result, even though increasing K is
discovered to always improve or at least maintain the performance.
| stat |
A multi-series framework for demand forecasts in E-commerce | Sales forecasts are crucial for the E-commerce business. State-of-the-art
techniques typically apply only univariate methods to make prediction for each
series independently. However, due to the short nature of sales times series in
E-commerce, univariate methods don't apply well. In this article, we propose a
global model which outperforms state-of-the-art models on real dataset. It is
achieved by using Tree Boosting Methods that exploit non-linearity and
cross-series information. We also proposed a preprocessing framework to
overcome the inherent difficulties in the E-commerce data. In particular, we
use different schemes to limit the impact of the volatility of the data.
| stat |
Exploring the Predictability of Cryptocurrencies via Bayesian Hidden
Markov Models | In this paper, we consider a variety of multi-state Hidden Markov models for
predicting and explaining the Bitcoin, Ether and Ripple returns in the presence
of state (regime) dynamics. In addition, we examine the effects of several
financial, economic and cryptocurrency specific predictors on the
cryptocurrency return series. Our results indicate that the Non-Homogeneous
Hidden Markov (NHHM) model with four states has the best one-step-ahead
forecasting performance among all competing models for all three series. The
dominance of the predictive densities over the single regime random walk model
relies on the fact that the states capture alternating periods with distinct
return characteristics. In particular, the four state NHHM model distinguishes
bull, bear and calm regimes for the Bitcoin series, and periods with different
profit and risk magnitudes for the Ether and Ripple series. Also, conditionally
on the hidden states, it identifies predictors with different linear and
non-linear effects on the cryptocurrency returns. These empirical findings
provide important insight for portfolio management and policy implementation.
| stat |
Factor Analysis for High-Dimensional Time Series with Change Point | We consider change-point latent factor models for high-dimensional time
series, where a structural break may exist in the underlying factor structure.
In particular, we propose consistent estimators for factor loading spaces
before and after the change point, and the problem of estimating the
change-point location is also considered. Compared with existing results on
change-point factor analysis of high-dimensional time series, a distinguished
feature of the current paper is that our results allow strong cross-sectional
dependence in the noise process. To accommodate the unknown degree of
cross-sectional dependence strength, we propose to use self-normalization to
pivotalize the change-point test statistic. Numerical experiments including a
Monte Carlo simulation study and a real data application are presented to
illustrate the proposed methods.
| stat |
The conditionally autoregressive hidden Markov model (CarHMM): Inferring
behavioural states from animal tracking data exhibiting conditional
autocorrelation | One of the central interests of animal movement ecology is relating movement
characteristics to behavioural characteristics. The traditional discrete-time
statistical tool for inferring unobserved behaviours from movement data is the
hidden Markov model (HMM). While the HMM is an important and powerful tool,
sometimes it is not flexible enough to appropriately fit the data. Data for
marine animals often exhibit conditional autocorrelation, self-dependence of
the step length process which cannot be explained solely by the behavioural
state, which violates one of the main assumptions of the HMM. Using a grey seal
track as an example, along with multiple simulation scenarios, we motivate and
develop the conditionally autoregressive hidden Markov model (CarHMM), which is
a generalization of the HMM designed specifically to handle conditional
autocorrelation.
In addition to introducing and examining the new CarHMM, we provide
guidelines for all stages of an analysis using either an HMM or CarHMM. These
include guidelines for pre-processing location data to obtain deflection angles
and step lengths, model selection, and model checking. In addition to these
practical guidelines, we link estimated model parameters to biologically
meaningful quantities such as activity budget and residency time. We also
provide interpretations of traditional "foraging" and "transiting" behaviours
in the context of the new CarHMM parameters.
| stat |
Panel Experiments and Dynamic Causal Effects: A Finite Population
Perspective | In panel experiments, we randomly expose multiple units to different
interventions and measure their subsequent outcomes, sequentially repeating the
procedure numerous times. Using the potential outcomes framework, we define
finite population dynamic causal effects that capture the relative
effectiveness of alternative treatment paths. For the leading example, known as
the lag-p dynamic causal effects, we provide a nonparametric estimator that is
unbiased over the randomization distribution. We then derive the finite
population limiting distribution of our estimators as either the sample size or
the duration of the experiment increases. Our approach provides a new technique
for deriving finite population central limit theorems that exploits the
underlying Martingale property of unbiased estimators. We further describe two
methods for conducting inference on dynamic causal effects: a conservative test
for weak null hypotheses of zero average causal effects using the limiting
distribution and an exact randomization-based test for sharp null hypotheses.
We also derive the finite population probability limit of commonly-used linear
fixed effects estimators, showing that these estimators perform poorly in the
presence of dynamic causal effects. We conclude with a simulation study and an
empirical application in which we reanalyze a lab experiment on cooperation.
| stat |
A general kernel boosting framework integrating pathways for predictive
modeling based on genomic data | Predictive modeling based on genomic data has gained popularity in biomedical
research and clinical practice by allowing researchers and clinicians to
identify biomarkers and tailor treatment decisions more efficiently. Analysis
incorporating pathway information can boost discovery power and better connect
new findings with biological mechanisms. In this article, we propose a general
framework, Pathway-based Kernel Boosting (PKB), which incorporates clinical
information and prior knowledge about pathways for prediction of binary,
continuous and survival outcomes. We introduce appropriate loss functions and
optimization procedures for different outcome types. Our prediction algorithm
incorporates pathway knowledge by constructing kernel function spaces from the
pathways and use them as base learners in the boosting procedure. Through
extensive simulations and case studies in drug response and cancer survival
datasets, we demonstrate that PKB can substantially outperform other competing
methods, better identify biological pathways related to drug response and
patient survival, and provide novel insights into cancer pathogenesis and
treatment response.
| stat |
Enabling Counterfactual Survival Analysis with Balanced Representations | Balanced representation learning methods have been applied successfully to
counterfactual inference from observational data. However, approaches that
account for survival outcomes are relatively limited. Survival data are
frequently encountered across diverse medical applications, i.e., drug
development, risk profiling, and clinical trials, and such data are also
relevant in fields like manufacturing (e.g., for equipment monitoring). When
the outcome of interest is a time-to-event, special precautions for handling
censored events need to be taken, as ignoring censored outcomes may lead to
biased estimates. We propose a theoretically grounded unified framework for
counterfactual inference applicable to survival outcomes. Further, we formulate
a nonparametric hazard ratio metric for evaluating average and individualized
treatment effects. Experimental results on real-world and semi-synthetic
datasets, the latter of which we introduce, demonstrate that the proposed
approach significantly outperforms competitive alternatives in both
survival-outcome prediction and treatment-effect estimation.
| stat |
Sequential estimation for GEE with adaptive variables and subject
selection | Modeling correlated or highly stratified multiple-response data becomes a
common data analysis task due to modern data monitoring facilities and methods.
Generalized estimating equations (GEE) is one of the popular statistical
methods for analyzing this kind of data. In this paper, we present a sequential
estimation procedure for obtaining GEE-based estimates. In addition to the
conventional random sampling, the proposed method features adaptive subject
recruiting and variable selection. Moreover, we equip our method with an
adaptive shrinkage property so that it can decide the effective variables
during the estimation procedure and build a confidence set with a pre-specified
precision for the corresponding parameters. In addition to the statistical
properties of the proposed procedure, we assess our method using both simulated
data and real data sets.
| stat |
Density ratio model with data-adaptive basis function | In many applications, we collect independent samples from interconnected
populations. These population distributions share some latent structure, so it
is advantageous to jointly analyze the samples. One effective way to connect
the distributions is the semiparametric density ratio model (DRM). A key
ingredient in the DRM is that the log density ratios are linear combinations of
prespecified functions; the vector formed by these functions is called the
basis function. A sensible basis function can often be chosen based on
knowledge of the context, and DRM-based inference is effective even if the
basis function is imperfect. However, a data-adaptive approach to the choice of
basis function remains an interesting and important research problem. We
propose an approach based on the classical functional principal component
analysis (FPCA). Under some conditions, we show that this approach leads to
consistent basis function estimation. Our simulation results show that the
proposed adaptive choice leads to an efficiency gain. We use a real-data
example to demonstrate the efficiency gain and the ease of our approach.
| stat |
Learning Integral Representations of Gaussian Processes | We propose a representation of Gaussian processes (GPs) based on powers of
the integral operator defined by a kernel function, we call these stochastic
processes integral Gaussian processes (IGPs). Sample paths from IGPs are
functions contained within the reproducing kernel Hilbert space (RKHS) defined
by the kernel function, in contrast sample paths from the standard GP are not
functions within the RKHS. We develop computationally efficient non-parametric
regression models based on IGPs. The main innovation in our regression
algorithm is the construction of a low dimensional subspace that captures the
information most relevant to explaining variation in the response. We use ideas
from supervised dimension reduction to compute this subspace. The result of
using the construction we propose involves significant improvements in the
computational complexity of estimating kernel hyper-parameters as well as
reducing the prediction variance.
| stat |
Characterizing functional relationships between anthropogenic and
biological sounds: A western New York state soundscape case study | Roads are a widespread feature of landscapes worldwide, and road traffic
sound potentially makes nearby habitat unsuitable for acoustically
communicating organisms. It is important to understand the influence of roads
at the soundscape level to mitigate negative impacts of road sound on
individual species as well as subsequent effects on the surrounding landscape.
We seek to characterize the relationship between anthropogenic and biological
sounds in western New York and assess the extent to which available traffic
data explains variability in anthropogenic noise. Recordings were obtained in
the spring of 2016 at 18 sites throughout western New York. We used the Welch
Power Spectral Density (PSD) at low frequencies (0.5-2 kHz) to represent
anthropogenic noise and PSD values at higher frequencies (2-11 kHz) to
represent biological sound. Relationships were modeled using a novel two-stage
hierarchical Bayesian model utilizing beta regression and basis splines. Model
results and map predictions illustrate that anthropogenic noise and biological
sound have an inverse relationship, and anthropogenic noise is greatest in
close proximity to high traffic volume roads. The predictions have large
uncertainty, resulting from the temporal coarseness of public road data used as
a proxy for traffic sound. Results suggest that finer temporal resolution
traffic sound data, such as crowd-sourced time-indexed traffic data from
geographic positioning systems, might better account for observed temporal
changes in the soundscape. The use of such data, in combination with the
proposed modeling framework, could have important implications for the
development of sound management policies.
| stat |
Flexible Modeling of Hurdle Conway-Maxwell-Poisson Distributions with
Application to Mining Injuries | While the hurdle Poisson regression is a popular class of models for count
data with excessive zeros, the link function in the binary component may be
unsuitable for highly imbalanced cases. Ordinary Poisson regression is unable
to handle the presence of dispersion. In this paper, we introduce
Conway-Maxwell-Poisson (CMP) distribution and integrate use of flexible skewed
Weibull link functions as better alternative. We take a fully Bayesian approach
to draw inference from the underlying models to better explain skewness and
quantify dispersion, with Deviance Information Criteria (DIC) used for model
selection. For empirical investigation, we analyze mining injury data for
period 2013-2016 from the U.S. Mine Safety and Health Administration (MSHA).
The risk factors describing proportions of employee hours spent in each type of
mining work are compositional data; the probabilistic principal components
analysis (PPCA) is deployed to deal with such covariates. The hurdle CMP
regression is additionally adjusted for exposure, measured by the total
employee working hours, to make inference on rate of mining injuries; we tested
its competitiveness against other models. This can be used as predictive model
in the mining workplace to identify features that increase the risk of injuries
so that prevention can be implemented.
| stat |
What is meant by 'P(R|Yobs)'? | Missing at Random (MAR) is a central concept in incomplete data methods, and
often it is stated as $P(R\mspace{3mu}|\,Y_{obs}, Y_{mis}) =
P(R\mspace{3mu}|\,Y_{obs})$. This notation has been used in the literature for
more than three decades and has become the de facto standard. In some cases,
the notation has been misinterpreted to be a statement about conditional
independence. While previous work has sought to clarify the required
definitions, a clear explanation of how to interpret the standard notation is
lacking, and a definition of the function $P(R\mspace{3mu}|\,Y_{obs})$ for
non-MAR mechanisms is difficult to locate in the literature. The aim of this
paper is to fill these gaps.
| stat |
A stochastic model for the lifecycle and track of extreme extratropical
cyclones in the North Atlantic | Extratropical cyclones are large-scale weather systems which are often the
source of extreme weather events in Northern Europe, often leading to mass
infrastructural damage and casualties. Such systems create a local vorticity
maxima which tracks across the Atlantic Ocean and from which can be determined
a climatology for the region. While there have been considerable advances in
developing algorithms for extracting the track and evolution of cyclones from
reanalysis datasets, the data record is relatively short. This justifies the
need for a statistical model to represent the more extreme characteristics of
these weather systems, specifically their intensity and the spatial variability
in their tracks. This paper presents a novel simulation-based approach to
modelling the lifecycle of extratropical cyclones in terms of both their tracks
and vorticity, incorporating various aspects of cyclone evolution and movement.
By drawing on methods from extreme value analysis, we can simulate more extreme
storms than those observed, representing a useful tool for practitioners
concerned with risk assessment with regard to these weather systems.
| stat |
Robust subset selection | The best subset selection (or "best subsets") estimator is a classic tool for
sparse regression, and developments in mathematical optimization over the past
decade have made it more computationally tractable than ever. Notwithstanding
its desirable statistical properties, the best subsets estimator is susceptible
to outliers and can break down in the presence of a single contaminated data
point. To address this issue, we propose a robust adaption of best subsets that
is highly resistant to contamination in both the response and the predictors.
Our estimator generalizes the notion of subset selection to both predictors and
observations, thereby achieving robustness in addition to sparsity. This
procedure, which we call "robust subset selection" (or "robust subsets"), is
defined by a combinatorial optimization problem for which we apply modern
discrete optimization methods. We formally establish the robustness of our
estimator in terms of the finite-sample breakdown point of its objective value.
In support of this result, we report experiments on both synthetic and real
data that demonstrate the superiority of robust subsets over best subsets in
the presence of contamination. Importantly, robust subsets fares competitively
across several metrics compared with popular robust adaptions of the Lasso.
| stat |
A Bayesian Approach for Predicting Food and Beverage Sales in Staff
Canteens and Restaurants | Accurate demand forecasting is one of the key aspects for successfully
managing restaurants and staff canteens. In particular, properly predicting
future sales of menu items allows a precise ordering of food stock. From an
environmental point of view, this ensures maintaining a low level of
pre-consumer food waste, while from the managerial point of view, this is
critical to guarantee the profitability of the restaurant. Hence, we are
interested in predicting future values of the daily sold quantities of given
menu items. The corresponding time series show multiple strong seasonalities,
trend changes, data gaps, and outliers. We propose a forecasting approach that
is solely based on the data retrieved from Point of Sales systems and allows
for a straightforward human interpretation. Therefore, we propose two
generalized additive models for predicting the future sales. In an extensive
evaluation, we consider two data sets collected at a casual restaurant and a
large staff canteen consisting of multiple time series, that cover a period of
20 months, respectively. We show that the proposed models fit the features of
the considered restaurant data. Moreover, we compare the predictive performance
of our method against the performance of other well-established forecasting
approaches.
| stat |
Adaptive Estimation in Structured Factor Models with Applications to
Overlapping Clustering | This work introduces a novel estimation method, called LOVE, of the entries
and structure of a loading matrix A in a sparse latent factor model X = AZ + E,
for an observable random vector X in Rp, with correlated unobservable factors Z
\in RK, with K unknown, and independent noise E. Each row of A is scaled and
sparse. In order to identify the loading matrix A, we require the existence of
pure variables, which are components of X that are associated, via A, with one
and only one latent factor. Despite the fact that the number of factors K, the
number of the pure variables, and their location are all unknown, we only
require a mild condition on the covariance matrix of Z, and a minimum of only
two pure variables per latent factor to show that A is uniquely defined, up to
signed permutations. Our proofs for model identifiability are constructive, and
lead to our novel estimation method of the number of factors and of the set of
pure variables, from a sample of size n of observations on X. This is the first
step of our LOVE algorithm, which is optimization-free, and has low
computational complexity of order p2. The second step of LOVE is an easily
implementable linear program that estimates A. We prove that the resulting
estimator is minimax rate optimal up to logarithmic factors in p. The model
structure is motivated by the problem of overlapping variable clustering,
ubiquitous in data science. We define the population level clusters as groups
of those components of X that are associated, via the sparse matrix A, with the
same unobservable latent factor, and multi-factor association is allowed.
Clusters are respectively anchored by the pure variables, and form overlapping
sub-groups of the p-dimensional random vector X. The Latent model approach to
OVErlapping clustering is reflected in the name of our algorithm, LOVE.
| stat |
Estimation of Accurate and Calibrated Uncertainties in Deterministic
models | In this paper we focus on the problem of assigning uncertainties to
single-point predictions generated by a deterministic model that outputs a
continuous variable. This problem applies to any state-of-the-art physics or
engineering models that have a computational cost that does not readily allow
to run ensembles and to estimate the uncertainty associated to single-point
predictions. Essentially, we devise a method to easily transform a
deterministic prediction into a probabilistic one. We show that for doing so,
one has to compromise between the accuracy and the reliability (calibration) of
such a probabilistic model. Hence, we introduce a cost function that encodes
their trade-off. We use the Continuous Rank Probability Score to measure
accuracy and we derive an analytic formula for the reliability, in the case of
forecasts of continuous scalar variables expressed in terms of Gaussian
distributions. The new Accuracy-Reliability cost function is then used to
estimate the input-dependent variance, given a black-box mean function, by
solving a two-objective optimization problem. The simple philosophy behind this
strategy is that predictions based on the estimated variances should not only
be accurate, but also reliable (i.e. statistical consistent with observations).
Conversely, early works based on the minimization of classical cost functions,
such as the negative log probability density, cannot simultaneously enforce
both accuracy and reliability. We show several examples both with synthetic
data, where the underlying hidden noise can accurately be recovered, and with
large real-world datasets.
| stat |
Defending the P-value | Attacks on the P-value are nothing new, but the recent attacks are
increasingly more serious. They come from more mainstream sources, with
widening targets such as a call to retire the significance testing altogether.
While well meaning, I believe these attacks are nevertheless misdirected:
Blaming the P-value for the naturally tentative trial-and-error process of
scientific discoveries, and presuming that banning the P-value would make the
process cleaner and less error-prone. However tentative, the skeptical
scientists still have to form unambiguous opinions, proximately to move forward
in their investigations and ultimately to present results to the wider
community. With obvious reasons, they constantly need to balance between the
false-positive and false-negative errors. How would banning the P-value or
significance tests help in this balancing act? It seems trite to say that this
balance will always depend on the relative costs or the trade-off between the
errors. These costs are highly context specific, varying by area of
applications or by stage of investigation. A calibrated but tunable knob, such
as that given by the P-value, is needed for controlling this balance. This
paper presents detailed arguments in support of the P-value.
| stat |
Robust Extrinsic Regression Analysis for Manifold Valued Data | Recently, there has been a growing need in analyzing data on manifolds owing
to their important role in diverse fields of science and engineering. In the
literature of manifold-valued data analysis up till now, however, only a few
works have been carried out concerning the robustness of estimation against
noises, outliers, and other sources of perturbations. In this regard, we
introduce a novel extrinsic framework for analyzing manifold valued data in a
robust manner. First, by extending the notion of the geometric median, we
propose a new robust location parameter on manifolds, so-called the extrinsic
median. A robust extrinsic regression method is also developed by incorporating
the conditional extrinsic median into the classical local polynomial regression
method. We present the Weiszfeld's algorithm for implementing the proposed
methods. The promising performance of our approach against existing methods is
illustrated through simulation studies.
| stat |
Identification and Estimation of Causal Effects from Dependent Data | The assumption that data samples are independent and identically distributed
(iid) is standard in many areas of statistics and machine learning.
Nevertheless, in some settings, such as social networks, infectious disease
modeling, and reasoning with spatial and temporal data, this assumption is
false. An extensive literature exists on making causal inferences under the iid
assumption [18, 12, 28, 22], even when unobserved confounding bias may be
present. But, as pointed out in [20], causal inference in non-iid contexts is
challenging due to the presence of both unobserved confounding and data
dependence. In this paper we develop a general theory describing when causal
inferences are possible in such scenarios. We use segregated graphs [21], a
generalization of latent projection mixed graphs [30], to represent causal
models of this type and provide a complete algorithm for non-parametric
identification in these models. We then demonstrate how statistical inference
may be performed on causal parameters identified by this algorithm. In
particular, we consider cases where only a single sample is available for parts
of the model due to full interference, i.e., all units are pathwise dependent
and neighbors' treatments affect each others' outcomes [26]. We apply these
techniques to a synthetic data set which considers users sharing fake news
articles given the structure of their social network, user activity levels, and
baseline demographics and socioeconomic covariates.
| stat |
On the Anatomy of MCMC-Based Maximum Likelihood Learning of Energy-Based
Models | This study investigates the effects of Markov chain Monte Carlo (MCMC)
sampling in unsupervised Maximum Likelihood (ML) learning. Our attention is
restricted to the family of unnormalized probability densities for which the
negative log density (or energy function) is a ConvNet. We find that many of
the techniques used to stabilize training in previous studies are not
necessary. ML learning with a ConvNet potential requires only a few
hyper-parameters and no regularization. Using this minimal framework, we
identify a variety of ML learning outcomes that depend solely on the
implementation of MCMC sampling.
On one hand, we show that it is easy to train an energy-based model which can
sample realistic images with short-run Langevin. ML can be effective and stable
even when MCMC samples have much higher energy than true steady-state samples
throughout training. Based on this insight, we introduce an ML method with
purely noise-initialized MCMC, high-quality short-run synthesis, and the same
budget as ML with informative MCMC initialization such as CD or PCD. Unlike
previous models, our energy model can obtain realistic high-diversity samples
from a noise signal after training.
On the other hand, ConvNet potentials learned with non-convergent MCMC do not
have a valid steady-state and cannot be considered approximate unnormalized
densities of the training data because long-run MCMC samples differ greatly
from observed images. We show that it is much harder to train a ConvNet
potential to learn a steady-state over realistic images. To our knowledge,
long-run MCMC samples of all previous models lose the realism of short-run
samples. With correct tuning of Langevin noise, we train the first ConvNet
potentials for which long-run and steady-state MCMC samples are realistic
images.
| stat |
Depth-based Weighted Jackknife Empirical Likelihood for Non-smooth
U-structure Equations | In many applications, parameters of interest are estimated by solving some
non-smooth estimating equations with $U$-statistic structure. Jackknife
empirical likelihood (JEL) approach can solve this problem efficiently by
reducing the computation complexity of the empirical likelihood (EL) method.
However, as EL, JEL suffers the sensitivity problem to outliers. In this paper,
we propose a weighted jackknife empirical likelihood (WJEL) to tackle the above
limitation of JEL. The proposed WJEL tilts the JEL function by assigning
smaller weights to outliers. The asymptotic of the WJEL ratio statistic is
derived. It converges in distribution to a multiple of a chi-square random
variable. The multiplying constant depends on the weighting scheme. The
self-normalized version of WJEL ratio does not require to know the constant and
hence yields the standard chi-square distribution in the limit. Robustness of
the proposed method is illustrated by simulation studies and one real data
application.
| stat |
Discovering Drug-Drug and Drug-Disease Interactions Inducing Acute
Kidney Injury Using Deep Rule Forests | Patients with Acute Kidney Injury (AKI) increase mortality, morbidity, and
long-term adverse events. Therefore, early identification of AKI may improve
renal function recovery, decrease comorbidities, and further improve patients'
survival. To control certain risk factors and develop targeted prevention
strategies are important to reduce the risk of AKI. Drug-drug interactions and
drug-disease interactions are critical issues for AKI. Typical statistical
approaches cannot handle the complexity of drug-drug and drug-disease
interactions. In this paper, we propose a novel learning algorithm, Deep Rule
Forests (DRF), which discovers rules from multilayer tree models as the
combinations of drug usages and disease indications to help identify such
interactions. We found that several disease and drug usages are considered
having significant impact on the occurrence of AKI. Our experimental results
also show that the DRF model performs comparatively better than typical
tree-based and other state-of-the-art algorithms in terms of prediction
accuracy and model interpretability.
| stat |
Bayesian Optimization of Risk Measures | We consider Bayesian optimization of objective functions of the form $\rho[
F(x, W) ]$, where $F$ is a black-box expensive-to-evaluate function and $\rho$
denotes either the VaR or CVaR risk measure, computed with respect to the
randomness induced by the environmental random variable $W$. Such problems
arise in decision making under uncertainty, such as in portfolio optimization
and robust systems design. We propose a family of novel Bayesian optimization
algorithms that exploit the structure of the objective function to
substantially improve sampling efficiency. Instead of modeling the objective
function directly as is typical in Bayesian optimization, these algorithms
model $F$ as a Gaussian process, and use the implied posterior on the objective
function to decide which points to evaluate. We demonstrate the effectiveness
of our approach in a variety of numerical experiments.
| stat |
Simulation-Based Decision Making in the NFL using NFLSimulatoR | In this paper, we introduce an R software package for simulating plays and
drives using play-by-play data from the National Football League. The
simulations are generated by sampling play-by-play data from previous football
seasons.The sampling procedure adds statistical rigor to any decisions or
inferences arising from examining the simulations. We highlight that the
package is particularly useful as a data-driven tool for evaluating potential
in-game strategies or rule changes within the league. We demonstrate its
utility by evaluating the oft-debated strategy of $\textit{going for it}$ on
fourth down and investigating whether or not teams should pass more than the
current standard.
| stat |
Fair Regression with Wasserstein Barycenters | We study the problem of learning a real-valued function that satisfies the
Demographic Parity constraint. It demands the distribution of the predicted
output to be independent of the sensitive attribute. We consider the case that
the sensitive attribute is available for prediction. We establish a connection
between fair regression and optimal transport theory, based on which we derive
a close form expression for the optimal fair predictor. Specifically, we show
that the distribution of this optimum is the Wasserstein barycenter of the
distributions induced by the standard regression function on the sensitive
groups. This result offers an intuitive interpretation of the optimal fair
prediction and suggests a simple post-processing algorithm to achieve fairness.
We establish risk and distribution-free fairness guarantees for this procedure.
Numerical experiments indicate that our method is very effective in learning
fair models, with a relative increase in error rate that is inferior to the
relative gain in fairness.
| stat |
On projection methods for functional time series forecasting | Two nonparametric methods are presented for forecasting functional time
series (FTS). The FTS we observe is a curve at a discrete-time point. We
address both one-step-ahead forecasting and dynamic updating. Dynamic updating
is a forward prediction of the unobserved segment of the most recent curve.
Among the two proposed methods, the first one is a straightforward adaptation
to FTS of the $k$-nearest neighbors methods for univariate time series
forecasting. The second one is based on a selection of curves, termed \emph{the
curve envelope}, that aims to be representative in shape and magnitude of the
most recent functional observation, either a whole curve or the observed part
of a partially observed curve. In a similar fashion to $k$-nearest neighbors
and other projection methods successfully used for time series forecasting, we
``project'' the $k$-nearest neighbors and the curves in the envelope for
forecasting. In doing so, we keep track of the next period evolution of the
curves. The methods are applied to simulated data, daily electricity demand,
and NOx emissions and provide competitive results with and often superior to
several benchmark predictions. The approach offers a model-free alternative to
statistical methods based on FTS modeling to study the cyclic or seasonal
behavior of many FTS.
| stat |
Markov Neighborhood Regression for High-Dimensional Inference | This paper proposes an innovative method for constructing confidence
intervals and assessing p-values in statistical inference for high-dimensional
linear models. The proposed method has successfully broken the high-dimensional
inference problem into a series of low-dimensional inference problems: For each
regression coefficient $\beta_i$, the confidence interval and $p$-value are
computed by regressing on a subset of variables selected according to the
conditional independence relations between the corresponding variable $X_i$ and
other variables. Since the subset of variables forms a Markov neighborhood of
$X_i$ in the Markov network formed by all the variables $X_1,X_2,\ldots,X_p$,
the proposed method is coined as Markov neighborhood regression. The proposed
method is tested on high-dimensional linear, logistic and Cox regression. The
numerical results indicate that the proposed method significantly outperforms
the existing ones. Based on the Markov neighborhood regression, a method of
learning causal structures for high-dimensional linear models is proposed and
applied to identification of drug sensitive genes and cancer driver genes. The
idea of using conditional independence relations for dimension reduction is
general and potentially can be extended to other high-dimensional or big data
problems as well.
| stat |
Blending of Probability and Non-Probability Samples: Applications to a
Survey of Military Caregivers | Probability samples are the preferred method for providing inferences that
are generalizable to a larger population. However, when a small (or rare)
subpopulation is the group of interest, this approach is unlikely to yield a
sample size large enough to produce precise inferences. Non-probability (or
convenience) sampling often provides the necessary sample size to yield
efficient estimates, but selection bias may compromise the generalizability of
results to the broader population. Motivating the exposition is a survey of
military caregivers; our interest is focused on unpaid caregivers of wounded,
ill, or injured servicemembers and veterans who served in the US armed forces
following September 11, 2001. An extensive probability sampling effort yielded
only 72 caregivers from this subpopulation. Therefore, we consider
supplementing the probability sample with a convenience sample from the same
subpopulation, and we develop novel methods of statistical weighting that may
be used to combine (or blend) the samples. Our analyses show that the
subpopulation of interest endures greater hardships than caregivers of veterans
with earlier dates of service, and these conclusions are discernably stronger
when blended samples with the proposed weighting schemes are used. We conclude
with simulation studies that illustrate the efficacy of the proposed
techniques, examine the bias-variance trade-off encountered when using
inadequately blended data, and show that the gain in precision provided by the
convenience sample is lower in circumstances where the outcome is strongly
related to the auxiliary variables used for blending.
| stat |
Adaptive-to-model hybrid of tests for regressions | In model checking for regressions, nonparametric estimation-based tests
usually have tractable limiting null distributions and are sensitive to
oscillating alternative models, but suffer from the curse of dimensionality. In
contrast, empirical process-based tests can, at the fastest possible rate,
detect local alternatives distinct from the null model, but is less sensitive
to oscillating alternative models and with intractable limiting null
distributions. It has long been an issue on how to construct a test that can
fully inherit the merits of these two types of tests and avoid the
shortcomings. We in this paper propose a generic adaptive-to-model hybrid of
moment and conditional moment-based test to achieve this goal. Further, a
significant feature of the method is to make nonparametric estimation-based
tests, under the alternatives, also share the merits of existing empirical
process-based tests. This methodology can be readily applied to other kinds of
data and constructing other hybrids. As a by-product in sufficient dimension
reduction field, the estimation of residual-related central subspace is used to
indicate the underlying models for model adaptation. A systematic study is
devoted to showing when alternative models can be indicated and when cannot.
This estimation is of its own interest and can be applied to the problems with
other kinds of data. Numerical studies are conducted to verify the powerfulness
of the proposed test.
| stat |
Self-driving car safety quantification via component-level analysis | In this paper, we present a rigorous modular statistical approach for arguing
safety or its insufficiency of an autonomous vehicle through a concrete
illustrative example. The methodology relies on making appropriate quantitative
studies of the performance of constituent components. We explain the importance
of sufficient and necessary conditions at the component level for the overall
safety of the vehicle as well as the cost-saving benefits of the approach. A
simple concrete automated braking example studied illustrates how separate
perception system and operational design domain statistical analyses can be
used to prove or disprove safety at the vehicle level.
| stat |
Smoothing Graphons for Modelling Exchangeable Relational Data | Modelling exchangeable relational data can be described by \textit{graphon
theory}. Most Bayesian methods for modelling exchangeable relational data can
be attributed to this framework by exploiting different forms of graphons.
However, the graphons adopted by existing Bayesian methods are either
piecewise-constant functions, which are insufficiently flexible for accurate
modelling of the relational data, or are complicated continuous functions,
which incur heavy computational costs for inference. In this work, we introduce
a smoothing procedure to piecewise-constant graphons to form {\em smoothing
graphons}, which permit continuous intensity values for describing relations,
but without impractically increasing computational costs. In particular, we
focus on the Bayesian Stochastic Block Model (SBM) and demonstrate how to adapt
the piecewise-constant SBM graphon to the smoothed version. We initially
propose the Integrated Smoothing Graphon (ISG) which introduces one smoothing
parameter to the SBM graphon to generate continuous relational intensity
values. We then develop the Latent Feature Smoothing Graphon (LFSG), which
improves on the ISG by introducing auxiliary hidden labels to decompose the
calculation of the ISG intensity and enable efficient inference. Experimental
results on real-world data sets validate the advantages of applying smoothing
strategies to the Stochastic Block Model, demonstrating that smoothing graphons
can greatly improve AUC and precision for link prediction without increasing
computational complexity.
| stat |
Analysing the causal effect of London cycle superhighways on traffic
congestion | Transport operators have a range of intervention options available to improve
or enhance their networks. Such interventions are often made in the absence of
sound evidence on resulting outcomes. Cycling superhighways were promoted as a
sustainable and healthy travel mode, one of the aims of which was to reduce
traffic congestion. Estimating the impacts that cycle superhighways have on
congestion is complicated due to the non-random assignment of such intervention
over the transport network. In this paper, we analyse the causal effect of
cycle superhighways utilising pre-intervention and post-intervention
information on traffic and road characteristics along with socio-economic
factors. We propose a modeling framework based on the propensity score and
outcome regression model. The method is also extended to the doubly robust
set-up. Simulation results show the superiority of the performance of the
proposed method over existing competitors. The method is applied to analyse a
real dataset on the London transport network. The methodology proposed can
assist in effective decision making to improve network performance.
| stat |
Importance sampling correction versus standard averages of reversible
MCMCs in terms of the asymptotic variance | We establish an ordering criterion for the asymptotic variances of two
consistent Markov chain Monte Carlo (MCMC) estimators: an importance sampling
(IS) estimator, based on an approximate reversible chain and subsequent IS
weighting, and a standard MCMC estimator, based on an exact reversible chain.
Essentially, we relax the criterion of the Peskun type covariance ordering by
considering two different invariant probabilities, and obtain, in place of a
strict ordering of asymptotic variances, a bound of the asymptotic variance of
IS by that of the direct MCMC. Simple examples show that IS can have
arbitrarily better or worse asymptotic variance than Metropolis-Hastings and
delayed-acceptance (DA) MCMC. Our ordering implies that IS is guaranteed to be
competitive up to a factor depending on the supremum of the (marginal) IS
weight. We elaborate upon the criterion in case of unbiased estimators as part
of an auxiliary variable framework. We show how the criterion implies
asymptotic variance guarantees for IS in terms of pseudo-marginal (PM) and DA
corrections, essentially if the ratio of exact and approximate likelihoods is
bounded. We also show that convergence of the IS chain can be less affected by
unbounded high-variance unbiased estimators than PM and DA chains.
| stat |
The Cost of Privacy: Optimal Rates of Convergence for Parameter
Estimation with Differential Privacy | Privacy-preserving data analysis is a rising challenge in contemporary
statistics, as the privacy guarantees of statistical methods are often achieved
at the expense of accuracy. In this paper, we investigate the tradeoff between
statistical accuracy and privacy in mean estimation and linear regression,
under both the classical low-dimensional and modern high-dimensional settings.
A primary focus is to establish minimax optimality for statistical estimation
with the $(\varepsilon,\delta)$-differential privacy constraint. To this end,
we find that classical lower bound arguments fail to yield sharp results, and
new technical tools are called for.
By refining the "tracing adversary" technique for lower bounds in the
theoretical computer science literature, we formulate a general lower bound
argument for minimax risks with differential privacy constraints, and apply
this argument to high-dimensional mean estimation and linear regression
problems. We also design computationally efficient algorithms that attain the
minimax lower bounds up to a logarithmic factor. In particular, for the
high-dimensional linear regression, a novel private iterative hard thresholding
pursuit algorithm is proposed, based on a privately truncated version of
stochastic gradient descent. The numerical performance of these algorithms is
demonstrated by simulation studies and applications to real data containing
sensitive information, for which privacy-preserving statistical methods are
necessary.
| stat |
A Hierarchical Max-Infinitely Divisible Spatial Model for Extreme
Precipitation | Understanding the spatial extent of extreme precipitation is necessary for
determining flood risk and adequately designing infrastructure (e.g.,
stormwater pipes) to withstand such hazards. While environmental phenomena
typically exhibit weakening spatial dependence at increasingly extreme levels,
limiting max-stable process models for block maxima have a rigid dependence
structure that does not capture this type of behavior. We propose a flexible
Bayesian model from a broader family of (conditionally) max-infinitely
divisible processes that allows for weakening spatial dependence at
increasingly extreme levels, and due to a hierarchical representation of the
likelihood in terms of random effects, our inference approach scales to large
datasets. Therefore, our model not only has a flexible dependence structure,
but it also allows for fast, fully Bayesian inference, prediction and
conditional simulation in high dimensions. The proposed model is constructed
using flexible random basis functions that are estimated from the data,
allowing for straightforward inspection of the predominant spatial patterns of
extremes. In addition, the described process possesses (conditional)
max-stability as a special case, making inference on the tail dependence class
possible. We apply our model to extreme precipitation in North-Eastern America,
and show that the proposed model adequately captures the extremal behavior of
the data. Interestingly, we find that the principal modes of spatial variation
estimated from our model resemble observed patterns in extreme precipitation
events occurring along the coast (e.g., with localized tropical cyclones and
convective storms) and mountain range borders. Our model, which can easily be
adapted to other types of environmental datasets, is therefore useful to
identify extreme weather patterns and regions at risk.
| stat |
Early warning in egg production curves from commercial hens: A SVM
approach | Artificial Intelligence allows the improvement of our daily life, for
instance, speech and handwritten text recognition, real time translation and
weather forecasting are common used applications. In the livestock sector,
machine learning algorithms have the potential for early detection and warning
of problems, which represents a significant milestone in the poultry industry.
Production problems generate economic loss that could be avoided by acting in a
timely manner. In the current study, training and testing of support vector
machines are addressed, for an early detection of problems in the production
curve of commercial eggs, using farm's egg production data of 478,919 laying
hens grouped in 24 flocks. Experiments using support vector machines with a 5
k-fold cross-validation were performed at different previous time intervals, to
alert with up to 5 days of forecasting interval, whether a flock will
experience a problem in production curve. Performance metrics such as accuracy,
specificity, sensitivity, and positive predictive value were evaluated,
reaching 0-day values of 0.9874, 0.9876, 0.9783 and 0.6518 respectively on
unseen data (test-set). The optimal forecasting interval was from zero to three
days, performance metrics decreases as the forecasting interval is increased.
It should be emphasized that this technique was able to issue an alert a day in
advance, achieving an accuracy of 0.9854, a specificity of 0.9865, a
sensitivity of 0.9333 and a positive predictive value of 0.6135. This novel
application embedded in a computer system of poultry management is able to
provide significant improvements in early detection and warning of problems
related to the production curve.
| stat |
A Statistical Analysis of Recent Traffic Crashes in Massachusetts | A statistical analysis implemented in the Python programming language was
performed on the available MassDOT car accident data to identify whether a
certain set of traffic circumstances would increase the likelihood of injuries.
In the analysis, we created a binary classifier as a model to separate crashes
that resulted in injury from those that did not. To accomplish this, we first
cleaned up the initial data, then proceeded to represent categorical variables
numerically through one hot encoding before finally producing models with
Recursive Feature Elimination (RFE) and without RFE, in conjunction with
logistic regression. This statistical analysis plays a significant role in our
modern road network that has presented us with a heap of obstacles, one of the
most critical being the issue of how we can ensure the safety of all drivers
and passengers. Findings from our analysis identify that tough weather and road
conditions, senior/teen drivers and dangerous intersections play prominent
roles in accidents that resulted in injuries in Massachusetts. These new
findings can provide valuable references and scientific data support to
relevant authorities and policy makers for upgrading road infrastructure,
passing regulations, etc.
| stat |
Estimation of Viterbi path in Bayesian hidden Markov models | The article studies different methods for estimating the Viterbi path in the
Bayesian framework. The Viterbi path is an estimate of the underlying state
path in hidden Markov models (HMMs), which has a maximum posterior probability
(MAP). For an HMM with given parameters, the Viterbi path can be easily found
with the Viterbi algorithm. In the Bayesian framework the Viterbi algorithm is
not applicable and several iterative methods can be used instead. We introduce
a new EM-type algorithm for finding the MAP path and compare it with various
other methods for finding the MAP path, including the variational Bayes
approach and MCMC methods. Examples with simulated data are used to compare the
performance of the methods. The main focus is on non-stochastic iterative
methods and our results show that the best of those methods work as well or
better than the best MCMC methods. Our results demonstrate that when the
primary goal is segmentation, then it is more reasonable to perform
segmentation directly by considering the transition and emission parameters as
nuisance parameters.
| stat |
Conditionally Independent Multiresolution Gaussian Processes | The multiresolution Gaussian process (GP) has gained increasing attention as
a viable approach towards improving the quality of approximations in GPs that
scale well to large-scale data. Most of the current constructions assume full
independence across resolutions. This assumption simplifies the inference, but
it underestimates the uncertainties in transitioning from one resolution to
another. This in turn results in models which are prone to overfitting in the
sense of excessive sensitivity to the chosen resolution, and predictions which
are non-smooth at the boundaries. Our contribution is a new construction which
instead assumes conditional independence among GPs across resolutions. We show
that relaxing the full independence assumption enables robustness against
overfitting, and that it delivers predictions that are smooth at the
boundaries. Our new model is compared against current state of the art on 2
synthetic and 9 real-world datasets. In most cases, our new conditionally
independent construction performed favorably when compared against models based
on the full independence assumption. In particular, it exhibits little to no
signs of overfitting.
| stat |
The Impact of Operating Environment on Efficiency of Public Libraries | Analysis of technical efficiency is an important tool in management of public
libraries. We assess the efficiency of 4660 public libraries established by
municipalities in the Czech Republic in the year 2017. For this purpose, we
utilize the data envelopment analysis (DEA) based on the Chebyshev distance. We
pay special attention to the operating environment and find that the efficiency
scores significantly depend on the population of the municipality and distance
to the municipality with extended powers. To remove the effect of the operating
environment, we perform DEA separately for categories based on the decision
tree analysis as well as categories designed by an expert.
| stat |
On Linear Stochastic Approximation: Fine-grained Polyak-Ruppert and
Non-Asymptotic Concentration | We undertake a precise study of the asymptotic and non-asymptotic properties
of stochastic approximation procedures with Polyak-Ruppert averaging for
solving a linear system $\bar{A} \theta = \bar{b}$. When the matrix $\bar{A}$
is Hurwitz, we prove a central limit theorem (CLT) for the averaged iterates
with fixed step size and number of iterations going to infinity. The CLT
characterizes the exact asymptotic covariance matrix, which is the sum of the
classical Polyak-Ruppert covariance and a correction term that scales with the
step size. Under assumptions on the tail of the noise distribution, we prove a
non-asymptotic concentration inequality whose main term matches the covariance
in CLT in any direction, up to universal constants. When the matrix $\bar{A}$
is not Hurwitz but only has non-negative real parts in its eigenvalues, we
prove that the averaged LSA procedure actually achieves an $O(1/T)$ rate in
mean-squared error. Our results provide a more refined understanding of linear
stochastic approximation in both the asymptotic and non-asymptotic settings. We
also show various applications of the main results, including the study of
momentum-based stochastic gradient methods as well as temporal difference
algorithms in reinforcement learning.
| stat |
Network Elastic Net for Identifying Smoking specific gene expression for
lung cancer | Survival month for non-small lung cancer patients depend upon which stage of
lung cancer is present. Our aim is to identify smoking specific gene expression
biomarkers in the prognosis of lung cancer patients. In this paper, we
introduce the network elastic net, a generalization of network lasso that
allows for simultaneous clustering and regression on graphs. In Network elastic
net, we consider similar patients based on smoking cigarettes per year to form
the network. We then further find the suitable cluster among patients based on
coefficients of genes having different survival month structures and showed the
efficacy of the clusters using stage enrichment. This can be used to identify
the stage of cancer using gene expression and smoking behavior of patients
without doing any tests.
| stat |
Flatness is a False Friend | Hessian based measures of flatness, such as the trace, Frobenius and spectral
norms, have been argued, used and shown to relate to generalisation. In this
paper we demonstrate that for feed forward neural networks under the cross
entropy loss, we would expect low loss solutions with large weights to have
small Hessian based measures of flatness. This implies that solutions obtained
using $L2$ regularisation should in principle be sharper than those without,
despite generalising better. We show this to be true for logistic regression,
multi-layer perceptrons, simple convolutional, pre-activated and wide residual
networks on the MNIST and CIFAR-$100$ datasets. Furthermore, we show that for
adaptive optimisation algorithms using iterate averaging, on the VGG-$16$
network and CIFAR-$100$ dataset, achieve superior generalisation to SGD but are
$30 \times$ sharper. This theoretical finding, along with experimental results,
raises serious questions about the validity of Hessian based sharpness measures
in the discussion of generalisation. We further show that the Hessian rank can
be bounded by the a constant times number of neurons multiplied by the number
of classes, which in practice is often a small fraction of the network
parameters. This explains the curious observation that many Hessian eigenvalues
are either zero or very near zero which has been reported in the literature.
| stat |
Fast Training of Sparse Graph Neural Networks on Dense Hardware | Graph neural networks have become increasingly popular in recent years due to
their ability to naturally encode relational input data and their ability to
scale to large graphs by operating on a sparse representation of graph
adjacency matrices. As we look to scale up these models using custom hardware,
a natural assumption would be that we need hardware tailored to sparse
operations and/or dynamic control flow. In this work, we question this
assumption by scaling up sparse graph neural networks using a platform targeted
at dense computation on fixed-size data. Drawing inspiration from optimization
of numerical algorithms on sparse matrices, we develop techniques that enable
training the sparse graph neural network model from Allamanis et al. [2018] in
13 minutes using a 512-core TPUv2 Pod, whereas the original training takes
almost a day.
| stat |
Combining Observational and Experimental Datasets Using Shrinkage
Estimators | We consider the problem of combining data from observational and experimental
sources to make causal conclusions. This problem is increasingly relevant, as
the modern era has yielded passive collection of massive observational datasets
in areas such as e-commerce and electronic health. These data may be used to
supplement experimental data, which is frequently expensive to obtain. In
Rosenman et al. (2018), we considered this problem under the assumption that
all confounders were measured. Here, we relax the assumption of
unconfoundedness. To derive combined estimators with desirable properties, we
make use of results from the Stein Shrinkage literature. Our contributions are
threefold. First, we propose a generic procedure for deriving shrinkage
estimators in this setting, making use of a generalized unbiased risk estimate.
Second, we develop two new estimators, prove finite sample conditions under
which they have lower risk than an estimator using only experimental data, and
show that each achieves a notion of asymptotic optimality. Third, we draw
connections between our approach and results in sensitivity analysis, including
proposing a method for evaluating the feasibility of our estimators.
| stat |
Non-parametric targeted Bayesian estimation of class proportions in
unlabeled data | We introduce a novel Bayesian estimator for the class proportion in an
unlabeled dataset, based on the targeted learning framework. Our procedure
requires the specification of a prior (and outputs a posterior) only for the
target of inference, instead of the prior (and posterior) on the full-data
distribution employed by classical non-parametric Bayesian methods .When the
scientific question can be characterized by a low-dimensional parameter
functional, focus on such a prior and posterior distributions is more aligned
with Bayesian subjectivism, compared to focus on entire data distributions. We
prove a Bernstein-von Mises-type result for our proposed Bayesian procedure,
which guarantees that the posterior distribution converges to the distribution
of an efficient, asymptotically linear estimator. In particular, the posterior
is Gaussian, doubly robust, and efficient in the limit, under the only
assumption that certain nuisance parameters are estimated at slow rates. We
perform numerical studies illustrating the frequentist properties of the
method. We also illustrate their use in a motivating application to estimate
the proportion of embolic strokes of undetermined source arising from occult
cardiac sources or large-artery atherosclerotic lesions. Though we focus on the
motivating example of the proportion of cases in an unlabeled dataset, the
procedure is general and can be adapted to estimate any pathwise differentiable
parameter in a non-parametric model.
| stat |
The GLD-plot: A depth-based plot to investigate unimodality of
directional data | A graphical tool for investigating unimodality of hyperspherical data is
proposed. It is based on the notion of statistical data depth function for
directional data which extends the univariate concept of rank. Firstly a local
version of distance-based depths for directional data based on aims at
analyzing the local structure of hyperspherical data is proposed. Then such
notion is compared to the global version of data depth by means of a
two-dimensional scatterplot, i.e. the GLD-plot. The proposal is illustrated on
simulated and real data examples.
| stat |
Designing Accurate Emulators for Scientific Processes using
Calibration-Driven Deep Models | Predictive models that accurately emulate complex scientific processes can
achieve exponential speed-ups over numerical simulators or experiments, and at
the same time provide surrogates for improving the subsequent analysis.
Consequently, there is a recent surge in utilizing modern machine learning (ML)
methods, such as deep neural networks, to build data-driven emulators. While
the majority of existing efforts has focused on tailoring off-the-shelf ML
solutions to better suit the scientific problem at hand, we study an often
overlooked, yet important, problem of choosing loss functions to measure the
discrepancy between observed data and the predictions from a model. Due to lack
of better priors on the expected residual structure, in practice, simple
choices such as the mean squared error and the mean absolute error are made.
However, the inherent symmetric noise assumption made by these loss functions
makes them inappropriate in cases where the data is heterogeneous or when the
noise distribution is asymmetric. We propose Learn-by-Calibrating (LbC), a
novel deep learning approach based on interval calibration for designing
emulators in scientific applications, that are effective even with
heterogeneous data and are robust to outliers. Using a large suite of
use-cases, we show that LbC provides significant improvements in generalization
error over widely-adopted loss function choices, achieves high-quality
emulators even in small data regimes and more importantly, recovers the
inherent noise structure without any explicit priors.
| stat |
Change point detection for COVID-19 excess deaths in Belgium | Emerging at the end of 2019, COVID-19 has become a public health threat to
people worldwide. Apart from the deaths who tested positive for COVID-19, many
others have died from causes indirectly related to COVID-19. Therefore, the
COVID-19 confirmed deaths underestimate the influence of the pandemic on the
society; instead, the measure of `excess deaths' is a more objective and
comparable way to assess the scale of the epidemic and formulate lessons. One
common practical issue in analyzing the impact of COVID-19 is to determine the
`pre-COVID-19' period and the `post-COVID-19' period. We apply a change point
detection method to identify any change points using the excess deaths in
Belgium.
| 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.