title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Ridge Regularizaton: an Essential Concept in Data Science | Ridge or more formally $\ell_2$ regularization shows up in many areas of
statistics and machine learning. It is one of those essential devices that any
good data scientist needs to master for their craft. In this brief ridge fest I
have collected together some of the magic and beauty of ridge that my
colleagues and I have encountered over the past 40 years in applied statistics.
| stat |
Meta-Learning Mean Functions for Gaussian Processes | When fitting Bayesian machine learning models on scarce data, the main
challenge is to obtain suitable prior knowledge and encode it into the model.
Recent advances in meta-learning offer powerful methods for extracting such
prior knowledge from data acquired in related tasks. When it comes to
meta-learning in Gaussian process models, approaches in this setting have
mostly focused on learning the kernel function of the prior, but not on
learning its mean function. In this work, we explore meta-learning the mean
function of a Gaussian process prior. We present analytical and empirical
evidence that mean function learning can be useful in the meta-learning
setting, discuss the risk of overfitting, and draw connections to other
meta-learning approaches, such as model agnostic meta-learning and functional
PCA.
| stat |
Vertex Classification on Weighted Networks | This paper proposes a discrimination technique for vertices in a weighted
network. We assume that the edge weights and adjacencies in the network are
conditionally independent and that both sources of information encode class
membership information. In particular, we introduce a edge weight distribution
matrix to the standard K-Block Stochastic Block Model to model weighted
networks. This allows us to develop simple yet powerful extensions of
classification techniques using the spectral embedding of the unweighted
adjacency matrix. We consider two assumptions on the edge weight distributions
and propose classification procedures in both settings. We show the
effectiveness of the proposed classifiers by comparing them to quadratic
discriminant analysis following the spectral embedding of a transformed
weighted network. Moreover, we discuss and show how the methods perform when
the edge weights do not encode class membership information.
| stat |
SEAM methodology for context-rich player matchup evaluations | We develop the SEAM (synthetic estimated average matchup) method for
describing batter versus pitcher matchups in baseball, both numerically and
visually. We first estimate the distribution of balls put into play by a batter
facing a pitcher, called the spray chart distribution. This distribution is
conditional on batter and pitcher characteristics. These characteristics are a
better expression of talent than any conventional statistics. Many individual
matchups have a sample size that is too small to be reliable. Synthetic
versions of the batter and pitcher under consideration are constructed in order
to alleviate these concerns. Weights governing how much influence these
synthetic players have on the overall spray chart distribution are constructed
to minimize expected mean square error. We then provide novel performance
metrics that are calculated as expectations taken with respect to the spray
chart distribution. These performance metrics provide a context rich approach
to player evaluation. Our main contribution is a Shiny app that allows users to
evaluate any batter-pitcher matchup that has occurred or could have occurred in
the last five years. One can access this app at
\url{https://seam.stat.illinois.edu/app/}. This interactive tool has utility
for anyone interested in baseball as well as team executives and players.
| stat |
Improved BiGAN training with marginal likelihood equalization | We propose a novel training procedure for improving the performance of
generative adversarial networks (GANs), especially to bidirectional GANs.
First, we enforce that the empirical distribution of the inverse inference
network matches the prior distribution, which favors the generator network
reproducibility on the seen samples. Second, we have found that the marginal
log-likelihood of the samples shows a severe overrepresentation of a certain
type of samples. To address this issue, we propose to train the bidirectional
GAN using a non-uniform sampling for the mini-batch selection, resulting in
improved quality and variety in generated samples measured quantitatively and
by visual inspection. We illustrate our new procedure with the well-known
CIFAR10, Fashion MNIST and CelebA datasets.
| stat |
An extended note on the multibin logarithmic score used in the FluSight
competitions | In recent years the Centers for Disease Control and Prevention (CDC) have
organized FluSight influenza forecasting competitions. To evaluate the
participants' forecasts a multibin logarithmic score has been created, which is
a non-standard variant of the established logarithmic score. Unlike the
original log score, the multibin version is not proper and may thus encourage
dishonest forecasting. We explore the practical consequences this may have,
using forecasts from the 2016/17 FluSight competition for illustration.
| stat |
Exploiting new forms of data to study the private rented sector:
strengths and limitations of a database of rental listings | Reviews of official statistics for UK housing have noted that developments
have not kept pace with real-world change, particularly the rapid growth of
private renting. This paper examines the potential value of big data in this
context. We report on the construction of a dataset from the on-line adverts of
one national lettings agency, describing the content of the dataset and efforts
to validate it against external sources. Focussing on one urban area, we
illustrate how the dataset can shed new light on local changes. Lastly, we
discuss the issues involved in making more routine use of this kind of data.
| stat |
A Bimodal Weibull Distribution: Properties and Inference | Modeling is a challenging topic and using parametric models is an important
stage to reach flexible function for modeling. Weibull distribution has two
parameters which are shape $\alpha$ and scale $\beta$. In this study,
bimodality parameter is added and so bimodal Weibull distribution is proposed
by using a quadratic transformation technique used to generate bimodal
functions produced due to using the quadratic expression. The analytical
simplicity of Weibull and quadratic form give an advantage to derive a bimodal
Weibull via constructing normalizing constant. The characteristics and
properties of the proposed distribution are examined to show its usability in
modeling. After examination as first stage in modeling issue, it is appropriate
to use bimodal Weibull for modeling data sets. Two estimation methods which are
maximum $\log_q$ likelihood and its special form including objective functions
$\log_q(f)$ and $\log(f)$ are used to estimate the parameters of shape, scale
and bimodality parameters of the function. The second stage in modeling is
overcome by using heuristic algorithm for optimization of function according to
parameters due to fact that converging to global point of objective function is
performed by heuristic algorithm based on the stochastic optimization. Real
data sets are provided to show the modeling competence of the proposed
distribution.
| stat |
A Variational Approach to Privacy and Fairness | In this article, we propose a new variational approach to learn private
and/or fair representations. This approach is based on the Lagrangians of a new
formulation of the privacy and fairness optimization problems that we propose.
In this formulation, we aim at generating representations of the data that keep
a prescribed level of the relevant information that is not shared by the
private or sensitive data, while minimizing the remaining information they
keep. The proposed approach (i) exhibits the similarities of the privacy and
fairness problems, (ii) allows us to control the trade-off between utility and
privacy or fairness through the Lagrange multiplier parameter, and (iii) can be
comfortably incorporated to common representation learning algorithms such as
the VAE, the $\beta$-VAE, the VIB, or the nonlinear IB.
| stat |
The shapes of an epidemic: using Functional Data Analysis to
characterize COVID-19 in Italy | We investigate patterns of COVID-19 mortality across 20 Italian regions and
their association with mobility, positivity, and socio-demographic,
infrastructural and environmental covariates. Notwithstanding limitations in
accuracy and resolution of the data available from public sources, we pinpoint
significant trends exploiting information in curves and shapes with Functional
Data Analysis techniques. These depict two starkly different epidemics; an
"exponential" one unfolding in Lombardia and the worst hit areas of the north,
and a milder, "flat(tened)" one in the rest of the country -- including Veneto,
where cases appeared concurrently with Lombardia but aggressive testing was
implemented early on. We find that mobility and positivity can predict COVID-19
mortality, also when controlling for relevant covariates. Among the latter,
primary care appears to mitigate mortality, and contacts in hospitals, schools
and work places to aggravate it. The techniques we describe could capture
additional and potentially sharper signals if applied to richer data.
| stat |
Kernel Methods for Unobserved Confounding: Negative Controls, Proxies,
and Instruments | Negative control is a strategy for learning the causal relationship between
treatment and outcome in the presence of unmeasured confounding. The treatment
effect can nonetheless be identified if two auxiliary variables are available:
a negative control treatment (which has no effect on the actual outcome), and a
negative control outcome (which is not affected by the actual treatment). These
auxiliary variables can also be viewed as proxies for a traditional set of
control variables, and they bear resemblance to instrumental variables. I
propose a new family of non-parametric algorithms for learning treatment
effects with negative controls. I consider treatment effects of the population,
of sub-populations, and of alternative populations. I allow for data that may
be discrete or continuous, and low-, high-, or infinite-dimensional. I impose
the additional structure of the reproducing kernel Hilbert space (RKHS), a
popular non-parametric setting in machine learning. I prove uniform consistency
and provide finite sample rates of convergence. I evaluate the estimators in
simulations.
| stat |
CytOpT: Optimal Transport with Domain Adaptation for Interpreting Flow
Cytometry data | The automated analysis of flow cytometry measurements is an active research
field. We introduce a new algorithm, referred to as CytOpT, using regularized
optimal transport to directly estimate the different cell population
proportions from a biological sample characterized with flow cytometry
measurements. We rely on the regularized Wasserstein metric to compare
cytometry measurements from different samples, thus accounting for possible
mis-alignment of a given cell population across sample (due to technical
variability from the technology of measurements). In this work, we rely on a
supervised learning technique based on the Wasserstein metric that is used to
estimate an optimal re-weighting of class proportions in a mixture model from a
source distribution (with known segmentation into cell sub-populations) to fit
a target distribution with unknown segmentation. Due to the high-dimensionality
of flow cytometry data, we use stochastic algorithms to approximate the
regularized Wasserstein metric to solve the optimization problem involved in
the estimation of optimal weights representing the cell population proportions
in the target distribution. Several flow cytometry data sets are used to
illustrate the performances of CytOpT that are also compared to those of
existing algorithms for automatic gating based on supervised learning.
| stat |
Optimizing Interim Analysis Timing for Bayesian Adaptive Commensurate
Designs | In developing products for rare diseases, statistical challenges arise due to
the limited number of patients available for participation in drug trials and
other clinical research. Bayesian adaptive clinical trial designs offer the
possibility of increased statistical efficiency, reduced development cost and
ethical hazard prevention via their incorporation of evidence from external
sources (historical data, expert opinions, and real-world evidence), and
flexibility in the specification of interim looks. In this paper, we propose a
novel Bayesian adaptive commensurate design that borrows adaptively from
historical information and also uses a particular payoff function to optimize
the timing of the study's interim analysis. The trial payoff is a function of
how many samples can be saved via early stopping and the probability of making
correct early decisions for either futility or efficacy. We calibrate our
Bayesian algorithm to have acceptable long-run frequentist properties (Type I
error and power) via simulation at the design stage. We illustrate our approach
using a pediatric trial design setting testing the effect of a new drug for a
rare genetic disease. The optimIA R package available at
https://github.com/wxwx1993/Bayesian_IA_Timing provides an easy-to-use
implementation of our approach.
| stat |
Foundations of Population-Based SHM, Part IV: The Geometry of Spaces of
Structures and their Feature Spaces | One of the requirements of the population-based approach to Structural Health
Monitoring (SHM) proposed in the earlier papers in this sequence, is that
structures be represented by points in an abstract space. Furthermore, these
spaces should be metric spaces in a loose sense; i.e. there should be some
measure of distance applicable to pairs of points; similar structures should
then be close in the metric. However, this geometrical construction is not
enough for the framing of problems in data-based SHM, as it leaves undefined
the notion of feature spaces. Interpreting the feature values on a
structure-by-structure basis as a type of field over the space of structures,
it seems sensible to borrow an idea from modern theoretical physics, and define
feature assignments as sections in a vector bundle over the structure space.
With this idea in place, one can interpret the effect of environmental and
operational variations as gauge degrees of freedom, as in modern gauge field
theories. This paper will discuss the various geometrical structures required
for an abstract theory of feature spaces in SHM, and will draw analogies with
how these structures have shown their power in modern physics. In the second
part of the paper, the problem of determining the normal condition cross
section of a feature bundle is addressed. The solution is provided by the
application of Graph Neural Networks (GNN), a versatile non-Euclidean machine
learning algorithm which is not restricted to inputs and outputs from vector
spaces. In particular, the algorithm is well suited to operating directly on
the sort of graph structures which are an important part of the proposed
framework for PBSHM. The solution of the normal section problem is demonstrated
for a heterogeneous population of truss structures for which the feature of
interest is the first natural frequency.
| stat |
Implicit differentiation of Lasso-type models for hyperparameter
optimization | Setting regularization parameters for Lasso-type estimators is notoriously
difficult, though crucial in practice. The most popular hyperparameter
optimization approach is grid-search using held-out validation data.
Grid-search however requires to choose a predefined grid for each parameter,
which scales exponentially in the number of parameters. Another approach is to
cast hyperparameter optimization as a bi-level optimization problem, one can
solve by gradient descent. The key challenge for these methods is the
estimation of the gradient with respect to the hyperparameters. Computing this
gradient via forward or backward automatic differentiation is possible yet
usually suffers from high memory consumption. Alternatively implicit
differentiation typically involves solving a linear system which can be
prohibitive and numerically unstable in high dimension. In addition, implicit
differentiation usually assumes smooth loss functions, which is not the case
for Lasso-type problems. This work introduces an efficient implicit
differentiation algorithm, without matrix inversion, tailored for Lasso-type
problems. Our approach scales to high-dimensional data by leveraging the
sparsity of the solutions. Experiments demonstrate that the proposed method
outperforms a large number of standard methods to optimize the error on
held-out data, or the Stein Unbiased Risk Estimator (SURE).
| stat |
Unified estimation framework for unnormalized models with statistical
efficiency | The parameter estimation of unnormalized models is a challenging problem. The
maximum likelihood estimation (MLE) is computationally infeasible for these
models since normalizing constants are not explicitly calculated. Although some
consistent estimators have been proposed earlier, the problem of statistical
efficiency remains. In this study, we propose a unified, statistically
efficient estimation framework for unnormalized models and several efficient
estimators, whose asymptotic variance is the same as the MLE. The computational
cost of these estimators is also reasonable and they can be employed whether
the sample space is discrete or continuous. The loss functions of the proposed
estimators are derived by combining the following two methods: (1)
density-ratio matching using Bregman divergence, and (2) plugging-in
nonparametric estimators. We also analyze the properties of the proposed
estimators when the unnormalized models are misspecified. The experimental
results demonstrate the advantages of our method over existing approaches.
| stat |
Stochastic Graph Recurrent Neural Network | Representation learning over graph structure data has been widely studied due
to its wide application prospects. However, previous methods mainly focus on
static graphs while many real-world graphs evolve over time. Modeling such
evolution is important for predicting properties of unseen networks. To resolve
this challenge, we propose SGRNN, a novel neural architecture that applies
stochastic latent variables to simultaneously capture the evolution in node
attributes and topology. Specifically, deterministic states are separated from
stochastic states in the iterative process to suppress mutual interference.
With semi-implicit variational inference integrated to SGRNN, a non-Gaussian
variational distribution is proposed to help further improve the performance.
In addition, to alleviate KL-vanishing problem in SGRNN, a simple and
interpretable structure is proposed based on the lower bound of KL-divergence.
Extensive experiments on real-world datasets demonstrate the effectiveness of
the proposed model. Code is available at
https://github.com/StochasticGRNN/SGRNN.
| stat |
Restricted Boltzmann Machines as Models of Interacting Variables | We study the type of distributions that Restricted Boltzmann Machines (RBMs)
with different activation functions can express by investigating the effect of
the activation function of the hidden nodes on the marginal distribution they
impose on observed binary nodes. We report an exact expression for these
marginals in the form of a model of interacting binary variables with the
explicit form of the interactions depending on the hidden node activation
function. We study the properties of these interactions in detail and evaluate
how the accuracy with which the RBM approximates distributions over binary
variables depends on the hidden node activation function and on the number of
hidden nodes. When the inferred RBM parameters are weak, an intuitive pattern
is found for the expression of the interaction terms which reduces
substantially the differences across activation functions. We show that the
weak parameter approximation is a good approximation for different RBMs trained
on the MNIST dataset. Interestingly, in these cases, the mapping reveals that
the inferred models are essentially low order interaction models.
| stat |
A Neural Process Approach for Probabilistic Reconstruction of No-Data
Gaps in Lunar Digital Elevation Maps | With the advent of NASA's lunar reconnaissance orbiter (LRO), a large amount
of high-resolution digital elevation maps (DEMs) have been constructed by using
narrow-angle cameras (NACs) to characterize the Moon's surface. However, NAC
DEMs commonly contain no-data gaps (voids), which makes the map less reliable.
To resolve the issue, this paper provides a deep-learning-based framework for
the probabilistic reconstruction of no-data gaps in NAC DEMs. The framework is
built upon a state of the art stochastic process model, attentive neural
processes (ANP), and predicts the conditional distribution of elevation on the
target coordinates (latitude and longitude) conditioned on the observed
elevation data in nearby regions. Furthermore, this paper proposes sparse
attentive neural processes (SANPs) that not only reduces the linear
computational complexity of the ANP O(N) to the constant complexity O(K) but
enhance the reconstruction performance by preventing overfitting and
over-smoothing problems. The proposed method is evaluated on the Apollo 17
landing site (20.0{\deg}N and 30.4{\deg}E), demonstrating that the suggested
approach successfully reconstructs no-data gaps with uncertainty analysis while
preserving the high resolution of original NAC DEMs.
| stat |
Is a single unique Bayesian network enough to accurately represent your
data? | Bayesian network (BN) modelling is extensively used in systems epidemiology.
Usually it consists in selecting and reporting the best-fitting structure
conditional to the data. A major practical concern is avoiding overfitting, on
account of its extreme flexibility and its modelling richness. Many approaches
have been proposed to control for overfitting. Unfortunately, they essentially
all rely on very crude decisions that result in too simplistic approaches for
such complex systems. In practice, with limited data sampled from complex
system, this approach seems too simplistic. An alternative would be to use the
Monte Carlo Markov chain model choice (MC3) over the network to learn the
landscape of reasonably supported networks, and then to present all possible
arcs with their MCMC support. This paper presents an R implementation, called
mcmcabn, of a flexible structural MC3 that is accessible to non-specialists.
| stat |
The Connection between Discrete- and Continuous-Time Descriptions of
Gaussian Continuous Processes | Learning the continuous equations of motion from discrete observations is a
common task in all areas of physics. However, not any discretization of a
Gaussian continuous-time stochastic process can be adopted in parametric
inference. We show that discretizations yielding consistent estimators have the
property of `invariance under coarse-graining', and correspond to fixed points
of a renormalization group map on the space of autoregressive moving average
(ARMA) models (for linear processes). This result explains why combining
differencing schemes for derivatives reconstruction and local-in-time inference
approaches does not work for time series analysis of second or higher order
stochastic differential equations, even if the corresponding integration
schemes may be acceptably good for numerical simulations.
| stat |
Model-Protected Multi-Task Learning | Multi-task learning (MTL) refers to the paradigm of learning multiple related
tasks together. In contrast, in single-task learning (STL) each individual task
is learned independently. MTL often leads to better trained models because they
can leverage the commonalities among related tasks. However, because MTL
algorithms can ``leak" information from different models across different
tasks, MTL poses a potential security risk. Specifically, an adversary may
participate in the MTL process through one task and thereby acquire the model
information for another task. The previously proposed privacy-preserving MTL
methods protect data instances rather than models, and some of them may
underperform in comparison with STL methods. In this paper, we propose a
privacy-preserving MTL framework to prevent information from each model leaking
to other models based on a perturbation of the covariance matrix of the model
matrix. We study two popular MTL approaches for instantiation, namely, learning
the low-rank and group-sparse patterns of the model matrix. Our algorithms can
be guaranteed not to underperform compared with STL methods. We build our
methods based upon tools for differential privacy, and privacy guarantees,
utility bounds are provided, and heterogeneous privacy budgets are considered.
The experiments demonstrate that our algorithms outperform the baseline methods
constructed by existing privacy-preserving MTL methods on the proposed
model-protection problem.
| stat |
Achieving Equalized Odds by Resampling Sensitive Attributes | We present a flexible framework for learning predictive models that
approximately satisfy the equalized odds notion of fairness. This is achieved
by introducing a general discrepancy functional that rigorously quantifies
violations of this criterion. This differentiable functional is used as a
penalty driving the model parameters towards equalized odds. To rigorously
evaluate fitted models, we develop a formal hypothesis test to detect whether a
prediction rule violates this property, the first such test in the literature.
Both the model fitting and hypothesis testing leverage a resampled version of
the sensitive attribute obeying equalized odds, by construction. We demonstrate
the applicability and validity of the proposed framework both in regression and
multi-class classification problems, reporting improved performance over
state-of-the-art methods. Lastly, we show how to incorporate techniques for
equitable uncertainty quantification---unbiased for each group under study---to
communicate the results of the data analysis in exact terms.
| stat |
Optimal stratification of survival data via Bayesian nonparametric
mixtures | The stratified proportional hazards model represents a simple solution to
account for heterogeneity within the data while keeping the multiplicative
effect on the hazard function. Strata are typically defined a priori by
resorting to the values taken by a categorical covariate. A general framework
is proposed, which allows for the stratification of a generic accelerated life
time model, including as a special case the Weibull proportional hazard model.
The stratification is determined a posteriori by taking into account that
strata might be characterized by different baseline survivals as well as
different effects of the predictors. This is achieved by considering a Bayesian
nonparametric mixture model and the posterior distribution it induces on the
space of data partitions. The optimal stratification is then identified by
means of the variation of information criterion and, in turn, stratum-specific
inference is carried out. The performance of the proposed method and its
robustness to the presence of right-censored observations are investigated by
means of an extensive simulation study. A further illustration is provided by
the analysis of a data set extracted from the University of Massachusetts AIDS
Research Unit IMPACT Study.
| stat |
Contrast Specific Propensity Scores | Basic propensity score methodology is designed to balance multivariate
pre-treatment covariates when comparing one active treatment with one control
treatment. Practical settings often involve comparing more than two treatments,
where more complicated contrasts than the basic treatment-control one,(1,-1),
are relevant. Here, we propose the use of contrast-specific propensity scores
(CSPS). CSPS allow the creation of treatment groups of units that are balanced
with respect to bifurcations of the specified contrasts and the multivariate
space spanned by them.
| stat |
Estimation, Confidence Intervals, and Large-Scale Hypotheses Testing for
High-Dimensional Mixed Linear Regression | This paper studies the high-dimensional mixed linear regression (MLR) where
the output variable comes from one of the two linear regression models with an
unknown mixing proportion and an unknown covariance structure of the random
covariates. Building upon a high-dimensional EM algorithm, we propose an
iterative procedure for estimating the two regression vectors and establish
their rates of convergence. Based on the iterative estimators, we further
construct debiased estimators and establish their asymptotic normality. For
individual coordinates, confidence intervals centered at the debiased
estimators are constructed.
Furthermore, a large-scale multiple testing procedure is proposed for testing
the regression coefficients and is shown to control the false discovery rate
(FDR) asymptotically. Simulation studies are carried out to examine the
numerical performance of the proposed methods and their superiority over
existing methods. The proposed methods are further illustrated through an
analysis of a dataset of multiplex image cytometry, which investigates the
interaction networks among the cellular phenotypes that include the expression
levels of 20 epitopes or combinations of markers.
| stat |
Discriminative Bayesian Filtering Lends Momentum to the Stochastic
Newton Method for Minimizing Log-Convex Functions | To minimize the average of a set of log-convex functions, the stochastic
Newton method iteratively updates its estimate using subsampled versions of the
full objective's gradient and Hessian. We contextualize this optimization
problem as sequential Bayesian inference on a latent state-space model with a
discriminatively-specified observation process. Applying Bayesian filtering
then yields a novel optimization algorithm that considers the entire history of
gradients and Hessians when forming an update. We establish matrix-based
conditions under which the effect of older observations diminishes over time,
in a manner analogous to Polyak's heavy ball momentum. We illustrate various
aspects of our approach with an example and review other relevant innovations
for the stochastic Newton method.
| stat |
Exploring the Effects of COVID-19 Containment Policies on Crime: An
Empirical Analysis of the Short-term Aftermath in Los Angeles | This work investigates whether and how COVID-19 containment policies had an
immediate impact on crime trends in Los Angeles. The analysis is conducted
using Bayesian structural time-series and focuses on nine crime categories and
on the overall crime count, daily monitored from January 1st 2017 to March 28th
2020. We concentrate on two post-intervention time windows - from March 4th to
March 16th and from March 4th to March 28th 2020 - to dynamically assess the
short-term effects of mild and strict policies. In Los Angeles, overall crime
has significantly decreased, as well as robbery, shoplifting, theft, and
battery. No significant effect has been detected for vehicle theft, burglary,
assault with a deadly weapon, intimate partner assault, and homicide. Results
suggest that, in the first weeks after the interventions are put in place,
social distancing impacts more directly on instrumental and less serious
crimes. Policy implications are also discussed.
| stat |
Multiple imputation with missing data indicators | Multiple imputation is a well-established general technique for analyzing
data with missing values. A convenient way to implement multiple imputation is
sequential regression multiple imputation (SRMI), also called chained equations
multiple imputation. In this approach, we impute missing values using
regression models for each variable, conditional on the other variables in the
data. This approach, however, assumes that the missingness mechanism is missing
at random, and it is not well-justified under not-at-random missingness without
additional modification. In this paper, we describe how we can generalize the
SRMI imputation procedure to handle not-at-random missingness (MNAR) in the
setting where missingness may depend on other variables that are also missing.
We provide algebraic justification for several generalizations of standard SRMI
using Taylor series and other approximations of the target imputation
distribution under MNAR. Resulting regression model approximations include
indicators for missingness, interactions, or other functions of the MNAR
missingness model and observed data. In a simulation study, we demonstrate that
the proposed SRMI modifications result in reduced bias in the final analysis
compared to standard SRMI, with an approximation strategy involving inclusion
of an offset in the imputation model performing the best overall. The method is
illustrated in a breast cancer study, where the goal is to estimate the
prevalence of a specific genetic pathogenic variant.
| stat |
Smoothly varying ridge regularization | A basis expansion with regularization methods is much appealing to the
flexible or robust nonlinear regression models for data with complex
structures. When the underlying function has inhomogeneous smoothness, it is
well known that conventional reguralization methods do not perform well. In
this case, an adaptive procedure such as a free-knot spline or a local
likelihood method is often introduced as an effective method. However, both
methods need intensive computational loads. In this study, we consider a new
efficient basis expansion by proposing a smoothly varying regularization method
which is constructed by some special penalties. We call them adaptive-type
penalties. In our modeling, adaptive-type penalties play key rolls and it has
been successful in giving good estimation for inhomogeneous smoothness
functions. A crucial issue in the modeling process is the choice of a suitable
model among candidates. To select the suitable model, we derive an approximated
generalized information criterion (GIC). The proposed method is investigated
through Monte Carlo simulations and real data analysis. Numerical results
suggest that our method performs well in various situations.
| stat |
HLIBCov: Parallel Hierarchical Matrix Approximation of Large Covariance
Matrices and Likelihoods with Applications in Parameter Identification | We provide more technical details about the HLIBCov package, which is using
parallel hierarchical ($\H$-) matrices to identify unknown parameters of the
covariance function (variance, smoothness, and covariance length). These
parameters are estimated by maximizing the joint Gaussian log-likelihood
function. The HLIBCov package approximates large dense inhomogeneous covariance
matrices with a log-linear computational cost and storage requirement. We
explain how to compute the Cholesky factorization, determinant, inverse and
quadratic form in the H-matrix format. To demonstrate the numerical
performance, we identify three unknown parameters in an example with 2,000,000
locations on a PC-desktop.
| stat |
Paired Comparisons Modeling using t-Distribution with Bayesian Analysis | A paired comparison analysis is the simplest way to make comparative
judgments between objects where objects may be goods, services or skills. For a
set of problems, this technique helps to choose the most important problem to
solve first and/or provides the solution that will be the most effective. This
paper presents the theory of paired comparisons method and contributes to the
paired comparisons models by developing a new model based on t-distribution.
The developed model is illustrated using a data set of citations among four
famous journals of Statistics. Using Bayesian analysis, the journals are ranked
as JRSS-B --> Biometrika --> JASA --> Comm. in Stats.
| stat |
Extrinsic Kernel Ridge Regression Classifier for Planar Kendall Shape
Space | Kernel methods have had great success in Statistics and Machine Learning.
Despite their growing popularity, however, less effort has been drawn towards
developing kernel based classification methods on Riemannian manifolds due to
difficulty in dealing with non-Euclidean geometry. In this paper, motivated by
the extrinsic framework of manifold-valued data analysis, we propose a new
positive definite kernel on planar Kendall shape space $\Sigma_2^k$, called
extrinsic Veronese Whitney Gaussian kernel. We show that our approach can be
extended to develop Gaussian kernels on any embedded manifold. Furthermore,
kernel ridge regression classifier (KRRC) is implemented to address the shape
classification problem on $\Sigma_2^k$, and their promising performances are
illustrated through the real data analysis.
| stat |
Conditional analysis for mixed covariates, with application to feed
intake of lactating sows | We propose a novel modeling framework to study the effect of covariates of
various types on the conditional distribution of the response. The methodology
accommodates flexible model structure, allows for joint estimation of the
quantiles at all levels, and involves a computationally efficient estimation
algorithm. Extensive numerical investigation confirms good performance of the
proposed method. The methodology is motivated by and applied to a lactating sow
study, where the primary interest is to understand how the dynamic change of
minute-by-minute temperature in the farrowing rooms within a day (functional
covariate) is associated with low quantiles of feed intake of lactating sows,
while accounting for other sow-specific information (vector covariate).
| stat |
Thirteen Simple Steps for Creating An R Package with an External C++
Library | We desribe how we extend R with an external C++ code library by using the
Rcpp package. Our working example uses the recent machine learning library and
application 'Corels' providing optimal yet easily interpretable rule lists
<arXiv:1704.01701> which we bring to R in the form of the 'RcppCorels' package.
We discuss each step in the process, and derive a set of simple rules and
recommendations which are illustrated with the concrete example.
| stat |
A Survey on Bayesian Deep Learning | A comprehensive artificial intelligence system needs to not only perceive the
environment with different `senses' (e.g., seeing and hearing) but also infer
the world's conditional (or even causal) relations and corresponding
uncertainty. The past decade has seen major advances in many perception tasks
such as visual object recognition and speech recognition using deep learning
models. For higher-level inference, however, probabilistic graphical models
with their Bayesian nature are still more powerful and flexible. In recent
years, Bayesian deep learning has emerged as a unified probabilistic framework
to tightly integrate deep learning and Bayesian models. In this general
framework, the perception of text or images using deep learning can boost the
performance of higher-level inference and in turn, the feedback from the
inference process is able to enhance the perception of text or images. This
survey provides a comprehensive introduction to Bayesian deep learning and
reviews its recent applications on recommender systems, topic models, control,
etc. Besides, we also discuss the relationship and differences between Bayesian
deep learning and other related topics such as Bayesian treatment of neural
networks. For a constantly updating project page, please refer to
https://github.com/js05212/BayesianDeepLearning-Survey.
| stat |
An Adaptive Algorithm based on High-Dimensional Function Approximation
to obtain Optimal Designs | Algorithms which compute locally optimal continuous designs often rely on a
finite design space or on repeatedly solving a complex non-linear program. Both
methods require extensive evaluations of the Jacobian Df of the underlying
model. These evaluations present a heavy computational burden. Based on the
Kiefer-Wolfowitz Equivalence Theorem we present a novel design of experiments
algorithm which computes optimal designs in a continuous design space. For this
iterative algorithm we combine an adaptive Bayes-like sampling scheme with
Gaussian process regression to approximate the directional derivative of the
design criterion. The approximation allows us to adaptively select new design
points on which to evaluate the model. The adaptive selection of the algorithm
requires significantly less evaluations of Df and reduces the runtime of the
computations. We show the viability of the new algorithm on two examples from
chemical engineering.
| stat |
A critical assessment of conformal prediction methods applied in binary
classification settings | In recent years there has been an increase in the number of scientific papers
that suggest using conformal predictions in drug discovery. We consider that
some versions of conformal predictions applied in binary settings are embroiled
in pitfalls, not obvious at first sight, and that it is important to inform the
scientific community about them. In the paper we first introduce the general
theory of conformal predictions and follow with an explanation of the versions
currently dominant in drug discovery research today. Finally, we provide cases
for their critical assessment in binary classification settings.
| stat |
Sample Size Estimation using a Latent Variable Model for Mixed Outcome
Co-Primary, Multiple Primary and Composite Endpoints | Mixed outcome endpoints that combine multiple continuous and discrete
components to form co-primary, multiple primary or composite endpoints are
often employed as primary outcome measures in clinical trials. There are many
advantages to joint modelling the individual outcomes using a latent variable
framework, however in order to make use of the model in practice we require
techniques for sample size estimation. In this paper we show how the latent
variable model can be applied to the three types of joint endpoints and propose
appropriate hypotheses, power and sample size estimation methods for each. We
illustrate the techniques using a numerical example based on the four
dimensional endpoint in the MUSE trial and find that the sample size required
for the co-primary endpoint is larger than that required for the individual
endpoint with the smallest effect size. Conversely, the sample size required
for the multiple primary endpoint is reduced from that required for the
individual outcome with the largest effect size. We show that the analytical
technique agrees with the empirical power from simulation studies. We further
illustrate the reduction in required sample size that may be achieved in trials
of mixed outcome composite endpoints through a simulation study and find that
the sample size primarily depends on the components driving response and the
correlation structure and much less so on the treatment effect structure in the
individual endpoints.
| stat |
Application of Levy Processes in Modelling (Geodetic) Time Series With
Mixed Spectra | Recently, various models have been developed, including the fractional
Brownian motion (fBm), to analyse the stochastic properties of geodetic time
series, together with the extraction of geophysical signals. The noise spectrum
of these time series is generally modeled as a mixed spectrum, with a sum of
white and coloured noise. Here, we are interested in modelling the residual
time series, after deterministically subtracting geophysical signals from the
observations. This residual time series is then assumed to be a sum of three
random variables (r.v.), with the last r.v. belonging to the family of Levy
processes. This stochastic term models the remaining residual signals and other
correlated processes. Via simulations and real time series, we identify three
classes of Levy processes: Gaussian, fractional and stable. In the first case,
residuals are predominantly constituted of short-memory processes. Fractional
Levy process can be an alternative model to the fBm in the presence of
long-term correlations and self-similarity property. Stable process is
characterized by a large variance, which can be satisfied in the case of
heavy-tailed distributions. The application to geodetic time series imply
potential anxiety in the functional model selection where missing geophysical
information can generate such residual time series.
| stat |
A confidence interval robust to publication bias for random-effects
meta-analysis of few studies | Systematic reviews aim to summarize all the available evidence relevant to a
particular research question. If appropriate, the data from identified studies
are quantitatively combined in a meta-analysis. Often only few studies
regarding a particular research question exist. In these settings the
estimation of the between-study heterogeneity is challenging. Furthermore, the
assessment of publication bias is difficult as standard methods such as visual
inspection or formal hypothesis tests in funnel plots do not provide adequate
guidance. Previously, Henmi and Copas (Statistics in Medicine 2010, 29:
2969--2983) proposed a confidence interval for the overall effect in
random-effects meta-analysis that is robust to publication bias to some extent.
As is evident from their simulations, the confidence intervals have improved
coverage compared with standard methods. To our knowledge, the properties of
their method has never been assessed for meta-analyses including fewer than
five studies. In this manuscript, we propose a variation of the method by Henmi
and Copas employing an improved estimator of the between-study heterogeneity,
in particular when dealing with few studies only. In a simulation study, the
proposed method is compared to several competitors. Overall, we found that our
method outperforms the others in terms of coverage probabilities. In
particular, an improvement compared with the proposal by Henmi and Copas is
demonstrated. The work is motivated and illustrated by a systematic review and
meta-analysis in paediatric immunosuppression following liver transplantations.
| stat |
Transport Monte Carlo: High-Accuracy Posterior Approximation via Random
Transport | In Bayesian applications, there is a huge interest in rapid and accurate
estimation of the posterior distribution, particularly for high dimensional or
hierarchical models. In this article, we propose to use optimization to solve
for a joint distribution (random transport plan) between two random variables,
$\theta$ from the posterior distribution and $\beta$ from the simple
multivariate uniform. Specifically, we obtain an approximate estimate of the
conditional distribution $\Pi(\beta\mid \theta)$ as an infinite mixture of
simple location-scale changes; applying the Bayes' theorem,
$\Pi(\theta\mid\beta)$ can be sampled as one of the reversed transforms from
the uniform, with the weight proportional to the posterior density/mass
function. This produces independent random samples with high approximation
accuracy, as well as nice theoretic guarantees. Our method shows compelling
advantages in performance and accuracy, compared to the state-of-the-art Markov
chain Monte Carlo and approximations such as variational Bayes and normalizing
flow. We illustrate this approach via several challenging applications, such as
sampling from multi-modal distribution, estimating sparse signals in high
dimension, and soft-thresholding of a graph with a prior on the degrees.
| stat |
Evaluating the performance of personal, social, health-related,
biomarker and genetic data for predicting an individuals future health using
machine learning: A longitudinal analysis | As we gain access to a greater depth and range of health-related information
about individuals, three questions arise: (1) Can we build better models to
predict individual-level risk of ill health? (2) How much data do we need to
effectively predict ill health? (3) Are new methods required to process the
added complexity that new forms of data bring? The aim of the study is to apply
a machine learning approach to identify the relative contribution of personal,
social, health-related, biomarker and genetic data as predictors of future
health in individuals. Using longitudinal data from 6830 individuals in the UK
from Understanding Society (2010-12 to 2015-17), the study compares the
predictive performance of five types of measures: personal (e.g. age, sex),
social (e.g. occupation, education), health-related (e.g. body weight, grip
strength), biomarker (e.g. cholesterol, hormones) and genetic single nucleotide
polymorphisms (SNPs). The predicted outcome variable was limiting long-term
illness one and five years from baseline. Two machine learning approaches were
used to build predictive models: deep learning via neural networks and XGBoost
(gradient boosting decision trees). Model fit was compared to traditional
logistic regression models. Results found that health-related measures had the
strongest prediction of future health status, with genetic data performing
poorly. Machine learning models only offered marginal improvements in model
accuracy when compared to logistic regression models, but also performed well
on other metrics e.g. neural networks were best on AUC and XGBoost on
precision. The study suggests that increasing complexity of data and methods
does not necessarily translate to improved understanding of the determinants of
health or performance of predictive models of ill health.
| stat |
Noisy-Input Entropy Search for Efficient Robust Bayesian Optimization | We consider the problem of robust optimization within the well-established
Bayesian optimization (BO) framework. While BO is intrinsically robust to noisy
evaluations of the objective function, standard approaches do not consider the
case of uncertainty about the input parameters. In this paper, we propose
Noisy-Input Entropy Search (NES), a novel information-theoretic acquisition
function that is designed to find robust optima for problems with both input
and measurement noise. NES is based on the key insight that the robust
objective in many cases can be modeled as a Gaussian process, however, it
cannot be observed directly. We evaluate NES on several benchmark problems from
the optimization literature and from engineering. The results show that NES
reliably finds robust optima, outperforming existing methods from the
literature on all benchmarks.
| stat |
Non-asymptotic error controlled sparse high dimensional precision matrix
estimation | Estimation of a high dimensional precision matrix is a critical problem to
many areas of statistics including Gaussian graphical models and inference on
high dimensional data. Working under the structural assumption of sparsity, we
propose a novel methodology for estimating such matrices while controlling the
false positive rate, percentage of matrix entries incorrectly chosen to be
non-zero. We specifically focus on false positive rates tending towards zero
with finite sample guarantees. This methodology is distribution free, but is
particularly applicable to the problem of Gaussian network recovery. We also
consider applications to constructing gene networks in genomics data.
| stat |
Distributed Variational Representation Learning | The problem of distributed representation learning is one in which multiple
sources of information $X_1,\ldots,X_K$ are processed separately so as to learn
as much information as possible about some ground truth $Y$. We investigate
this problem from information-theoretic grounds, through a generalization of
Tishby's centralized Information Bottleneck (IB) method to the distributed
setting. Specifically, $K$ encoders, $K \geq 2$, compress their observations
$X_1,\ldots,X_K$ separately in a manner such that, collectively, the produced
representations preserve as much information as possible about $Y$. We study
both discrete memoryless (DM) and memoryless vector Gaussian data models. For
the discrete model, we establish a single-letter characterization of the
optimal tradeoff between complexity (or rate) and relevance (or information)
for a class of memoryless sources (the observations $X_1,\ldots,X_K$ being
conditionally independent given $Y$). For the vector Gaussian model, we provide
an explicit characterization of the optimal complexity-relevance tradeoff.
Furthermore, we develop a variational bound on the complexity-relevance
tradeoff which generalizes the evidence lower bound (ELBO) to the distributed
setting. We also provide two algorithms that allow to compute this bound: i) a
Blahut-Arimoto type iterative algorithm which enables to compute optimal
complexity-relevance encoding mappings by iterating over a set of
self-consistent equations, and ii) a variational inference type algorithm in
which the encoding mappings are parametrized by neural networks and the bound
approximated by Markov sampling and optimized with stochastic gradient descent.
Numerical results on synthetic and real datasets are provided to support the
efficiency of the approaches and algorithms developed in this paper.
| stat |
Bayesian Model Selection for Change Point Detection and Clustering | We address the new problem of estimating a piece-wise constant signal with
the purpose of detecting its change points and the levels of clusters. Our
approach is to model it as a nonparametric penalized least square model
selection on a family of models indexed over the collection of partitions of
the design points and propose a computationally efficient algorithm to
approximately solve it. Statistically, minimizing such a penalized criterion
yields an approximation to the maximum a posteriori probability (MAP)
estimator. The criterion is then analyzed and an oracle inequality is derived
using a Gaussian concentration inequality. The oracle inequality is used to
derive on one hand conditions for consistency and on the other hand an adaptive
upper bound on the expected square risk of the estimator, which statistically
motivates our approximation. Finally, we apply our algorithm to simulated data
to experimentally validate the statistical guarantees and illustrate its
behavior.
| stat |
Winning by hiding behind others: An analysis of speed skating data | In some athletic races, such as cycling and types of speed skating races,
athletes have to complete a relatively long distance at a high speed in the
presence of direct opponents. To win such a race, athletes are motivated to
hide behind others to suppress energy consumption before a final moment of the
race. This situation seems to produce a social dilemma: players want to hide
behind others, whereas if a group of players attempts to do so, they may all
lose to other players that overtake them. To support that speed skaters are
involved in such a social dilemma, we analyzed video footage data for 14 mass
start skating races to find that skaters that hid behind others to avoid air
resistance for a long time before the final lap tended to win. Furthermore, the
finish rank of the skaters in mass start races was independent of the record of
the same skaters in time-trial races measured in the absence of direct
opponents. The results suggest that how to strategically cope with a skater's
dilemma may be a key determinant for winning long-distance and high-speed races
with direct opponents.
| stat |
A brief introduction to the Grey Machine Learning | This paper presents a brief introduction to the key points of the Grey
Machine Learning (GML) based on the kernels. The general formulation of the
grey system models have been firstly summarized, and then the nonlinear
extension of the grey models have been developed also with general
formulations. The kernel implicit mapping is used to estimate the nonlinear
function of the GML model, by extending the nonparametric formulation of the
LSSVM, the estimation of the nonlinear function of the GML model can also be
expressed by the kernels. A short discussion on the priority of this new
framework to the existing grey models and LSSVM have also been discussed in
this paper. And the perspectives and future orientations of this framework have
also been presented.
| stat |
Partially Pooled Propensity Score Models for Average Treatment Effect
Estimation with Multilevel Data | Causal inference analyses often use existing observational data, which in
many cases has some clustering of individuals. In this paper we discuss
propensity score weighting methods in a multilevel setting where within
clusters individuals share unmeasured confounders that are related to treatment
assignment and the potential outcomes. We focus in particular on settings where
models with fixed cluster effects are either not feasible or not useful due to
the presence of a large number of small clusters. We found, both through
numerical experiments and theoretical derivations, that a strategy of grouping
clusters with similar treatment prevalence and estimating propensity scores
within such cluster groups is effective in reducing bias from unmeasured
cluster-level covariates under mild conditions on the outcome model. We apply
our proposed method in evaluating the effectiveness of center-based pre-school
program participation on children's achievement at kindergarten, using the
Early Childhood Longitudinal Study, Kindergarten data.
| stat |
Covariate-assisted Sparse Tensor Completion | We aim to provably complete a sparse and highly-missing tensor in the
presence of covariate information along tensor modes. Our motivation comes from
online advertising where users click-through-rates (CTR) on ads over various
devices form a CTR tensor that has about 96% missing entries and has many zeros
on non-missing entries, which makes the standalone tensor completion method
unsatisfactory. Beside the CTR tensor, additional ad features or user
characteristics are often available. In this paper, we propose
Covariate-assisted Sparse Tensor Completion (COSTCO) to incorporate covariate
information for the recovery of the sparse tensor. The key idea is to jointly
extract latent components from both the tensor and the covariate matrix to
learn a synthetic representation. Theoretically, we derive the error bound for
the recovered tensor components and explicitly quantify the improvements on
both the reveal probability condition and the tensor recovery accuracy due to
covariates. Finally, we apply COSTCO to an advertisement dataset consisting of
a CTR tensor and ad covariate matrix, leading to 23% accuracy improvement over
the baseline. An important by-product is that ad latent components from COSTCO
reveal interesting ad clusters, which are useful for better ad targeting.
| stat |
On Posterior Consistency of Bayesian Factor Models in High Dimensions | As a principled dimension reduction technique, factor models have been widely
adopted in social science, economics, bioinformatics, and many other fields.
However, in high-dimensional settings, conducting a 'correct' Bayesianfactor
analysis can be subtle since it requires both a careful prescription of the
prior distribution and a suitable computational strategy. In particular, we
analyze the issues related to the attempt of being "noninformative" for
elements of the factor loading matrix, especially for sparse Bayesian factor
models in high dimensions, and propose solutions to them. We show here why
adopting the orthogonal factor assumption is appropriate and can result in a
consistent posterior inference of the loading matrix conditional on the true
idiosyncratic variance and the allocation of nonzero elements in the true
loading matrix. We also provide an efficient Gibbs sampler to conduct the full
posterior inference based on the prior setup from Rockova and George (2016)and
a uniform orthogonal factor assumption on the factor matrix.
| stat |
Cross-Validation and Uncertainty Determination for Randomized Neural
Networks with Applications to Mobile Sensors | Randomized artificial neural networks such as extreme learning machines
provide an attractive and efficient method for supervised learning under
limited computing ressources and green machine learning. This especially
applies when equipping mobile devices (sensors) with weak artificial
intelligence. Results are discussed about supervised learning with such
networks and regression methods in terms of consistency and bounds for the
generalization and prediction error. Especially, some recent results are
reviewed addressing learning with data sampled by moving sensors leading to
non-stationary and dependent samples.
As randomized networks lead to random out-of-sample performance measures, we
study a cross-validation approach to handle the randomness and make use of it
to improve out-of-sample performance. Additionally, a computationally efficient
approach to determine the resulting uncertainty in terms of a confidence
interval for the mean out-of-sample prediction error is discussed based on
two-stage estimation. The approach is applied to a prediction problem arising
in vehicle integrated photovoltaics.
| stat |
A Poisson multi-Bernoulli mixture filter for coexisting point and
extended targets | This paper proposes a Poisson multi-Bernoulli mixture (PMBM) filter for
coexisting point and extended targets. The PMBM filter provides a recursion to
compute the filtering posterior based on single-target predictions and updates.
In this paper, we first derive the PMBM filter update for a generalised
measurement model, which can include measurements originated from point and
extended targets. Second, we propose a single-target space that accommodates
both point and extended targets and derive the filtering recursion that
propagates Gaussian densities for single targets and gamma Gaussian inverse
Wishart densities for extended targets. As a computationally efficient
approximation of the PMBM filter, we also develop a Poisson multi-Bernoulli
(PMB) filter for coexisting point and extended targets. The resulting filters
are analysed via numerical simulations.
| stat |
Identifying brain hierarchical structures associated with Alzheimer's
disease using a regularized regression method with tree predictors | Brain segmentation at different levels is generally represented as
hierarchical trees. Brain regional atrophy at specific levels was found to be
marginally associated with Alzheimer's disease outcomes. In this study, we
propose an L1-type regularization for predictors that follow a hierarchical
tree structure. Considering a tree as a directed acyclic graph, we interpret
the model parameters from a path analysis perspective. Under this concept, the
proposed penalty regulates the total effect of each predictor on the outcome.
With regularity conditions, it is shown that under the proposed regularization,
the estimator of the model coefficient is consistent in L2-norm and the model
selection is also consistent. By applying to a brain structural magnetic
resonance imaging dataset acquired from the Alzheimer's Disease Neuroimaging
Initiative, the proposed approach identifies brain regions where atrophy in
these regions demonstrates the declination in memory. With regularization on
the total effects, the findings suggest that the impact of atrophy on memory
deficits is localized from small brain regions but at various levels of brain
segmentation.
| stat |
Nested Dirichlet Process For Population Size Estimation From Multi-list
Recapture Data | Heterogeneity of response patterns is important in estimating the size of a
closed population from multiple recapture data when capture patterns are
different over time and location. In this paper, we extend the non-parametric
one layer latent class model for multiple recapture data proposed by
Manrique-Vallier (2016) to a nested latent class model with the first layer
modeling individual heterogeneity and the second layer modeling location-time
differences. Location-time groups with similar recording patterns are in the
same top layer latent class and individuals within each top layer class are
dependent. The nested latent class model incorporates hierarchical
heterogeneity into the modeling to estimate population size from multi-list
recapture data. This approach leads to more accurate population size estimation
and reduced uncertainty. We apply the method to estimating casualties from the
Syrian conflict.
| stat |
Intrinsic wavelet regression for curves of Hermitian positive definite
matrices | Intrinsic wavelet transforms and wavelet estimation methods are introduced
for curves in the non-Euclidean space of Hermitian positive definite matrices,
with in mind the application to Fourier spectral estimation of multivariate
stationary time series. The main focus is on intrinsic average-interpolation
wavelet transforms in the space of positive definite matrices equipped with an
affine-invariant Riemannian metric, and convergence rates of linear wavelet
thresholding are derived for intrinsically smooth curves of Hermitian positive
definite matrices. In the context of multivariate Fourier spectral estimation,
intrinsic wavelet thresholding is equivariant under a change of basis of the
time series, and nonlinear wavelet thresholding is able to capture localized
features in the spectral density matrix across frequency, always guaranteeing
positive definite estimates. The finite-sample performance of intrinsic wavelet
thresholding is assessed by means of simulated data and compared to several
benchmark estimators in the Riemannian manifold. Further illustrations are
provided by examining the multivariate spectra of trial-replicated brain signal
time series recorded during a learning experiment.
| stat |
Graph Spectral Embedding for Parsimonious Transmission of Multivariate
Time Series | We propose a graph spectral representation of time series data that 1) is
parsimoniously encoded to user-demanded resolution; 2) is unsupervised and
performant in data-constrained scenarios; 3) captures event and
event-transition structure within the time series; and 4) has near-linear
computational complexity in both signal length and ambient dimension. This
representation, which we call Laplacian Events Signal Segmentation (LESS), can
be computed on time series of arbitrary dimension and originating from sensors
of arbitrary type. Hence, time series originating from sensors of heterogeneous
type can be compressed to levels demanded by constrained-communication
environments, before being fused at a common center.
Temporal dynamics of the data is summarized without explicit partitioning or
probabilistic modeling. As a proof-of-principle, we apply this technique on
high dimensional wavelet coefficients computed from the Free Spoken Digit
Dataset to generate a memory efficient representation that is interpretable.
Due to its unsupervised and non-parametric nature, LESS representations remain
performant in the digit classification task despite the absence of labels and
limited data.
| stat |
Probabilistic Solutions To Ordinary Differential Equations As Non-Linear
Bayesian Filtering: A New Perspective | We formulate probabilistic numerical approximations to solutions of ordinary
differential equations (ODEs) as problems in Gaussian process (GP) regression
with non-linear measurement functions. This is achieved by defining the
measurement sequence to consist of the observations of the difference between
the derivative of the GP and the vector field evaluated at the GP---which are
all identically zero at the solution of the ODE. When the GP has a state-space
representation, the problem can be reduced to a non-linear Bayesian filtering
problem and all widely-used approximations to the Bayesian filtering and
smoothing problems become applicable. Furthermore, all previous GP-based ODE
solvers that are formulated in terms of generating synthetic measurements of
the gradient field come out as specific approximations. Based on the non-linear
Bayesian filtering problem posed in this paper, we develop novel Gaussian
solvers for which we establish favourable stability properties. Additionally,
non-Gaussian approximations to the filtering problem are derived by the
particle filter approach. The resulting solvers are compared with other
probabilistic solvers in illustrative experiments.
| stat |
Estimating the number of communities in networks by spectral methods | Community detection is a fundamental problem in network analysis with many
methods available to estimate communities. Most of these methods assume that
the number of communities is known, which is often not the case in practice. We
study a simple and very fast method for estimating the number of communities
based on the spectral properties of certain graph operators, such as the
non-backtracking matrix and the Bethe Hessian matrix. We show that the method
performs well under several models and a wide range of parameters, and is
guaranteed to be consistent under several asymptotic regimes. We compare this
method to several existing methods for estimating the number of communities and
show that it is both more accurate and more computationally efficient.
| stat |
BreGMN: scaled-Bregman Generative Modeling Networks | The family of f-divergences is ubiquitously applied to generative modeling in
order to adapt the distribution of the model to that of the data.
Well-definedness of f-divergences, however, requires the distributions of the
data and model to overlap completely in every time step of training. As a
result, as soon as the support of distributions of data and model contain
non-overlapping portions, gradient based training of the corresponding model
becomes hopeless. Recent advances in generative modeling are full of remedies
for handling this support mismatch problem: key ideas include either modifying
the objective function to integral probability measures (IPMs) that are
well-behaved even on disjoint probabilities, or optimizing a well-behaved
variational lower bound instead of the true objective. We, on the other hand,
establish that a complete change of the objective function is unnecessary, and
instead an augmentation of the base measure of the problematic divergence can
resolve the issue. Based on this observation, we propose a generative model
which leverages the class of Scaled Bregman Divergences and generalizes both
f-divergences and Bregman divergences. We analyze this class of divergences and
show that with the appropriate choice of base measure it can resolve the
support mismatch problem and incorporate geometric information. Finally, we
study the performance of the proposed method and demonstrate promising results
on MNIST, CelebA and CIFAR-10 datasets.
| stat |
Practical Bayesian Modeling and Inference for Massive Spatial Datasets
On Modest Computing Environments | With continued advances in Geographic Information Systems and related
computational technologies, statisticians are often required to analyze very
large spatial datasets. This has generated substantial interest over the last
decade, already too vast to be summarized here, in scalable methodologies for
analyzing large spatial datasets. Scalable spatial process models have been
found especially attractive due to their richness and flexibility and,
particularly so in the Bayesian paradigm, due to their presence in hierarchical
model settings. However, the vast majority of research articles present in this
domain have been geared toward innovative theory or more complex model
development. Very limited attention has been accorded to approaches for easily
implementable scalable hierarchical models for the practicing scientist or
spatial analyst. This article is submitted to the Practice section of the
journal with the aim of developing massively scalable Bayesian approaches that
can rapidly deliver Bayesian inference on spatial process that are practically
indistinguishable from inference obtained using more expensive alternatives. A
key emphasis is on implementation within very standard (modest) computing
environments (e.g., a standard desktop or laptop) using easily available
statistical software packages without requiring message-parsing interfaces or
parallel programming paradigms. Key insights are offered regarding assumptions
and approximations concerning practical efficiency.
| stat |
Attention, Learn to Solve Routing Problems! | The recently presented idea to learn heuristics for combinatorial
optimization problems is promising as it can save costly development. However,
to push this idea towards practical implementation, we need better models and
better ways of training. We contribute in both directions: we propose a model
based on attention layers with benefits over the Pointer Network and we show
how to train this model using REINFORCE with a simple baseline based on a
deterministic greedy rollout, which we find is more efficient than using a
value function. We significantly improve over recent learned heuristics for the
Travelling Salesman Problem (TSP), getting close to optimal results for
problems up to 100 nodes. With the same hyperparameters, we learn strong
heuristics for two variants of the Vehicle Routing Problem (VRP), the
Orienteering Problem (OP) and (a stochastic variant of) the Prize Collecting
TSP (PCTSP), outperforming a wide range of baselines and getting results close
to highly optimized and specialized algorithms.
| stat |
Congestion in near capacity metro operations: optimum boardings and
alightings at bottleneck stations | During peak hours, metro systems often operate at high service frequencies to
transport large volumes of passengers. However, the punctuality of such
operations can be severely impacted by a vicious circle of passenger congestion
and train delays. In particular, high volumes of passenger boardings and
alightings may lead to increased dwell times at stations, that may eventually
cause queuing of trains in upstream. Such stations act as active bottlenecks in
the metro network and congestion may propagate from these bottlenecks to the
entire network. Thus, understanding the mechanism that drives passenger
congestion at these bottleneck stations is crucial to develop informed control
strategies, such as control of inflow of passengers entering these stations. To
this end, we conduct the first station-level econometric analysis to estimate a
causal relationship between boarding-alighting movements and train flow using
data from entry/exit gates and train movement data of the Mass Transit Railway,
Hong Kong. We adopt a Bayesian non-parametric spline-based regression approach
and apply instrumental variables estimation to control for confounding bias
that may occur due to unobserved characteristics of metro operations. Through
the results of the empirical study, we identify bottleneck stations and provide
estimates of optimum passenger movements per train and service frequencies at
the bottleneck stations. These estimates, along with real data on daily demand,
could assist metro operators in devising station-level control strategies.
| stat |
Multi-Output Gaussian Processes with Functional Data: A Study on Coastal
Flood Hazard Assessment | Most of the existing coastal flood Forecast and Early-Warning Systems do not
model the flood, but instead, rely on the prediction of hydrodynamic conditions
at the coast and on expert judgment. Recent scientific contributions are now
capable to precisely model flood events, even in situations where wave
overtopping plays a significant role. Such models are nevertheless
costly-to-evaluate and surrogate ones need to be exploited for substantial
computational savings. For the latter models, the hydro-meteorological forcing
conditions (inputs) or flood events (outputs) are conveniently parametrised
into scalar representations. However, they neglect the fact that inputs are
actually functions (more precisely, time series), and that floods spatially
propagate inland. Here, we introduce a multi-output Gaussian process model
accounting for both criteria. On various examples, we test its versatility for
both learning spatial maps and inferring unobserved ones. We demonstrate that
efficient implementations are obtained by considering tensor-structured data
and/or sparse-variational approximations. Finally, the proposed framework is
applied on a coastal application aiming at predicting flood events. We conclude
that accurate predictions are obtained in the order of minutes rather than the
couples of days required by dedicated hydrodynamic simulators.
| stat |
A Simple and Interpretable Predictive Model for Healthcare | Deep Learning based models are currently dominating most state-of-the-art
solutions for disease prediction. Existing works employ RNNs along with
multiple levels of attention mechanisms to provide interpretability. These deep
learning models, with trainable parameters running into millions, require huge
amounts of compute and data to train and deploy. These requirements are
sometimes so huge that they render usage of such models as unfeasible. We
address these challenges by developing a simpler yet interpretable non-deep
learning based model for application to EHR data. We model and showcase our
work's results on the task of predicting first occurrence of a diagnosis, often
overlooked in existing works. We push the capabilities of a tree based model
and come up with a strong baseline for more sophisticated models. Its
performance shows an improvement over deep learning based solutions (both, with
and without the first-occurrence constraint) all the while maintaining
interpretability.
| stat |
Cross-Entropy Based Importance Sampling for Stochastic Simulation Models | To efficiently evaluate system reliability based on Monte Carlo simulation,
importance sampling is used widely. The optimal importance sampling density was
derived in 1950s for the deterministic simulation model, which maps an input to
an output deterministically, and is approximated in practice using various
methods. For the stochastic simulation model whose output is random given an
input, the optimal importance sampling density was derived only recently. In
the existing literature, metamodel-based approaches have been used to
approximate this optimal density. However, building a satisfactory metamodel is
often difficult or time-consuming in practice. This paper proposes a
cross-entropy based method, which is automatic and does not require specific
domain knowledge. The proposed method uses an expectation-maximization
algorithm to guide the choice of a mixture distribution model for approximating
the optimal density. The method iteratively updates the approximated density to
minimize its estimated discrepancy, measured by estimated cross-entropy, from
the optimal density. The mixture model's complexity is controlled using the
cross-entropy information criterion. The method is empirically validated using
extensive numerical studies and applied to a case study of evaluating the
reliability of wind turbine using a stochastic simulation model.
| stat |
Gaussian process with derivative information for the analysis of the
sunlight adverse effects on color of rock art paintings | Microfading Spectrometry (MFS) is a method for assessing light sensitivity
color (spectral) variations of cultural heritage objects. The MFS technique
provides measurements of the surface under study, where each point of the
surface gives rise to a time-series that represents potential spectral (color)
changes due to sunlight exposition over time. Color fading is expected to be
non-decreasing as a function of time and stabilize eventually. These properties
can be expressed in terms of the partial derivatives of the functions. We
propose a spatio-temporal model that takes this information into account by
jointly modeling the spatio-temporal process and its derivative process using
Gaussian processes (GPs). We fitted the proposed model to MFS data collected
from the surface of prehistoric rock art paintings. A multivariate covariance
function in a GP allows modeling trichromatic image color variables jointly
with spatial distances and time points variables as inputs to evaluate the
covariance structure of the data. We demonstrated that the colorimetric
variables are useful for predicting the color fading time-series for new
unobserved spatial locations. Furthermore, constraining the model using
derivative sign observations for monotonicity was shown to be beneficial in
terms of both predictive performance and application-specific interpretability.
| stat |
Efficient identification of infected sub-population | When testing for infections, the standard method is to test each subject
individually. If testing methodology is such that samples from multiple
subjects can be efficiently combined and tested at once, yielding a positive
results if any one subject in the subgroup is positive, then one can often
identify the infected sub-population with a considerably lower number of tests
compared to the number of test subjects. We present two such methods that allow
an increase in testing efficiency (in terms of total number of test performed)
by a factor of $\approx$ 10 if population infection rate is $10^{-2}$ and a
factor of $\approx$50 when it is $10^{-3}$. Such methods could be useful when
testing large fractions of the total population, as will be perhaps required
during the current coronavirus pandemic.
| stat |
Model Evidence with Fast Tree Based Quadrature | High dimensional integration is essential to many areas of science, ranging
from particle physics to Bayesian inference. Approximating these integrals is
hard, due in part to the difficulty of locating and sampling from regions of
the integration domain that make significant contributions to the overall
integral. Here, we present a new algorithm called Tree Quadrature (TQ) that
separates this sampling problem from the problem of using those samples to
produce an approximation of the integral. TQ places no qualifications on how
the samples provided to it are obtained, allowing it to use state-of-the-art
sampling algorithms that are largely ignored by existing integration
algorithms. Given a set of samples, TQ constructs a surrogate model of the
integrand in the form of a regression tree, with a structure optimised to
maximise integral precision. The tree divides the integration domain into
smaller containers, which are individually integrated and aggregated to
estimate the overall integral. Any method can be used to integrate each
individual container, so existing integration methods, like Bayesian Monte
Carlo, can be combined with TQ to boost their performance. On a set of
benchmark problems, we show that TQ provides accurate approximations to
integrals in up to 15 dimensions; and in dimensions 4 and above, it outperforms
simple Monte Carlo and the popular Vegas method.
| stat |
Generalisation in fully-connected neural networks for time series
forecasting | In this paper we study the generalization capabilities of fully-connected
neural networks trained in the context of time series forecasting. Time series
do not satisfy the typical assumption in statistical learning theory of the
data being i.i.d. samples from some data-generating distribution. We use the
input and weight Hessians, that is the smoothness of the learned function with
respect to the input and the width of the minimum in weight space, to quantify
a network's ability to generalize to unseen data. While such generalization
metrics have been studied extensively in the i.i.d. setting of for example
image recognition, here we empirically validate their use in the task of time
series forecasting. Furthermore we discuss how one can control the
generalization capability of the network by means of the training process using
the learning rate, batch size and the number of training iterations as
controls. Using these hyperparameters one can efficiently control the
complexity of the output function without imposing explicit constraints.
| stat |
An axiomatic nonparametric production function estimator: Modeling
production in Japan's cardboard industry | We develop a new approach to estimate a production function based on the
economic axioms of the Regular Ultra Passum law and convex non-homothetic input
isoquants. Central to the development of our estimator is stating the axioms as
shape constraints and using shape constrained nonparametric regression methods.
We implement this approach using data from the Japanese corrugated cardboard
industry from 1997-2007. Using this new approach, we find most productive scale
size is a function of the capital-to-labor ratio and the largest firms operate
close to the largest most productive scale size associated with a high
capital-to-labor ratio. We measure the productivity growth across the panel
periods based on the residuals from our axiomatic model. We also decompose
productivity into scale, input mix, and unexplained effects to clarify the
sources the productivity differences and provide managers guidance to make
firms more productive.
| stat |
The problem of perfect predictors in statistical spike train models | Generalized Linear Models (GLMs) have been used extensively in statistical
models of spike train data. However, the IRLS algorithm, which is often used to
fit such models, can fail to converge in situations where response and
non-response can be separated by a single predictor or a linear combination of
multiple predictors. Such situations are likely to arise in many neural systems
due to properties such as refractoriness and incomplete sampling of the signals
that influence spiking. In this paper, we describe multiple classes of
approaches to address this problem: Standard IRLS with a fixed iteration limit,
computing the maximum likelihood solution in the limit, Bayesian estimation,
regularization, change of basis, and modifying the search parameters. We
demonstrate a specific application of each of these methods to spiking data
from rat somatosensory cortex and discuss the advantages and disadvantages of
each. We also provide an example of a roadmap for selecting a method based on
the problem's particular analysis issues and scientific goals.
| stat |
Metropolis-Hastings with Averaged Acceptance Ratios | Markov chain Monte Carlo (MCMC) methods to sample from a probability
distribution $\pi$ defined on a space $(\Theta,\mathcal{T})$ consist of the
simulation of realisations of Markov chains $\{\theta_{n},n\geq1\}$ of
invariant distribution $\pi$ and such that the distribution of $\theta_{i}$
converges to $\pi$ as $i\rightarrow\infty$. In practice one is typically
interested in the computation of expectations of functions, say $f$, with
respect to $\pi$ and it is also required that averages
$M^{-1}\sum_{n=1}^{M}f(\theta_{n})$ converge to the expectation of interest.
The iterative nature of MCMC makes it difficult to develop generic methods to
take advantage of parallel computing environments when interested in reducing
time to convergence. While numerous approaches have been proposed to reduce the
variance of ergodic averages, including averaging over independent realisations
of $\{\theta_{n},n\geq1\}$ simulated on several computers, techniques to reduce
the "burn-in" of MCMC are scarce. In this paper we explore a simple and generic
approach to improve convergence to equilibrium of existing algorithms which
rely on the Metropolis-Hastings (MH) update, the main building block of MCMC.
The main idea is to use averages of the acceptance ratio w.r.t. multiple
realisations of random variables involved, while preserving $\pi$ as invariant
distribution. The methodology requires limited change to existing code, is
naturally suited to parallel computing and is shown on our examples to provide
substantial performance improvements both in terms of convergence to
equilibrium and variance of ergodic averages. In some scenarios gains are
observed even on a serial machine.
| stat |
There is no trade-off: enforcing fairness can improve accuracy | One of the main barriers to the broader adoption of algorithmic fairness in
machine learning is the trade-off between fairness and performance of ML
models: many practitioners are unwilling to sacrifice the performance of their
ML model for fairness. In this paper, we show that this trade-off may not be
necessary. If the algorithmic biases in an ML model are due to sampling biases
in the training data, then enforcing algorithmic fairness may improve the
performance of the ML model on unbiased test data. We study conditions under
which enforcing algorithmic fairness helps practitioners learn the Bayes
decision rule for (unbiased) test data from biased training data. We also
demonstrate the practical implications of our theoretical results in real-world
ML tasks.
| stat |
Information-Theoretic Confidence Bounds for Reinforcement Learning | We integrate information-theoretic concepts into the design and analysis of
optimistic algorithms and Thompson sampling. By making a connection between
information-theoretic quantities and confidence bounds, we obtain results that
relate the per-period performance of the agent with its information gain about
the environment, thus explicitly characterizing the exploration-exploitation
tradeoff. The resulting cumulative regret bound depends on the agent's
uncertainty over the environment and quantifies the value of prior information.
We show applicability of this approach to several environments, including
linear bandits, tabular MDPs, and factored MDPs. These examples demonstrate the
potential of a general information-theoretic approach for the design and
analysis of reinforcement learning algorithms.
| stat |
Stochastic modeling of non-linear adsorption with Gaussian kernel
density estimators | Adsorption is a relevant process in many fields, such as product
manufacturing or pollution remediation in porous materials. Adsorption takes
place at the molecular scale, amenable to be modeled by Lagrangian numerical
methods. We have proposed a chemical diffusion-reaction model for the
simulation of adsorption, based on the combination of a random walk particle
tracking method involving the use of Gaussian Kernel Density Estimators. The
main feature of the proposed model is that it can effectively reproduce the
nonlinear behavior characteristic of the Langmuir and Freundlich isotherms. In
the former, it is enough to add a finite number of sorption sites of
homogeneous sorption properties, and to set the process as the combination of
the forward and the backward reactions, each one of them with a prespecified
reaction rate. To model the Freundlich isotherm instead, typical of low to
intermediate range of solute concentrations, there is a need to assign a
different equilibrium constant to each specific sorption site, provided they
are all drawn from a truncated power-law distribution. Both nonlinear models
can be combined in a single framework to obtain a typical observed behavior for
a wide range of concentration values.
| stat |
SIMEX Estimation in Parametric Modal Regression with Measurement Error | For a class of parametric modal regression models with measurement error, a
simulation extrapolation estimation procedure is proposed in this paper for
estimating the modal regression coefficients. Large sample properties of the
proposed estimation procedure, including the consistency and asymptotic
normality, are thoroughly investigated. Simulation studies are conducted to
evaluate its robustness to potential outliers and the effectiveness in reducing
the bias caused by the measurement error.
| stat |
Online Graph-Based Change-Point Detection for High Dimensional Data | Online change-point detection (OCPD) is important for application in various
areas such as finance, biology, and the Internet of Things (IoT). However, OCPD
faces major challenges due to high-dimensionality, and it is still rarely
studied in literature. In this paper, we propose a novel, online, graph-based,
change-point detection algorithm to detect change of distribution in low- to
high-dimensional data. We introduce a similarity measure, which is derived from
the graph-spanning ratio, to test statistically if a change occurs. Through
numerical study using artificial online datasets, our data-driven approach
demonstrates high detection power for high-dimensional data, while the false
alarm rate (type I error) is controlled at a nominal significant level. In
particular, our graph-spanning approach has desirable power with small and
multiple scanning window, which allows timely detection of change-point in the
online setting.
| stat |
BayesFlow: Learning complex stochastic models with invertible neural
networks | Estimating the parameters of mathematical models is a common problem in
almost all branches of science. However, this problem can prove notably
difficult when processes and model descriptions become increasingly complex and
an explicit likelihood function is not available. With this work, we propose a
novel method for globally amortized Bayesian inference based on invertible
neural networks which we call BayesFlow. The method uses simulation to learn a
global estimator for the probabilistic mapping from observed data to underlying
model parameters. A neural network pre-trained in this way can then, without
additional training or optimization, infer full posteriors on arbitrary many
real datasets involving the same model family. In addition, our method
incorporates a summary network trained to embed the observed data into
maximally informative summary statistics. Learning summary statistics from data
makes the method applicable to modeling scenarios where standard inference
techniques with hand-crafted summary statistics fail. We demonstrate the
utility of BayesFlow on challenging intractable models from population
dynamics, epidemiology, cognitive science and ecology. We argue that BayesFlow
provides a general framework for building amortized Bayesian parameter
estimation machines for any forward model from which data can be simulated.
| stat |
Local likelihood estimation of complex tail dependence structures,
applied to U.S. precipitation extremes | To disentangle the complex non-stationary dependence structure of
precipitation extremes over the entire contiguous U.S., we propose a flexible
local approach based on factor copula models. Our sub-asymptotic spatial
modeling framework yields non-trivial tail dependence structures, with a
weakening dependence strength as events become more extreme, a feature commonly
observed with precipitation data but not accounted for in classical asymptotic
extreme-value models. To estimate the local extremal behavior, we fit the
proposed model in small regional neighborhoods to high threshold exceedances,
under the assumption of local stationarity, which allows us to gain in
flexibility. Adopting a local censored likelihood approach, inference is made
on a fine spatial grid, and local estimation is performed by taking advantage
of distributed computing resources and the embarrassingly parallel nature of
this estimation procedure. The local model is efficiently fitted at all grid
points, and uncertainty is measured using a block bootstrap procedure. An
extensive simulation study shows that our approach can adequately capture
complex, non-stationary dependencies, while our study of U.S. winter
precipitation data reveals interesting differences in local tail structures
over space, which has important implications on regional risk assessment of
extreme precipitation events.
| stat |
Network Group Hawkes Process Model | In this work, we study the event occurrences of user activities on online
social network platforms. To characterize the social activity interactions
among network users, we propose a network group Hawkes (NGH) process model.
Particularly, the observed network structure information is employed to model
the users' dynamic posting behaviors. Furthermore, the users are clustered into
latent groups according to their dynamic behavior patterns. To estimate the
model, a constraint maximum likelihood approach is proposed. Theoretically, we
establish the consistency and asymptotic normality of the estimators. In
addition, we show that the group memberships can be identified consistently. To
conduct estimation, a branching representation structure is firstly introduced,
and a stochastic EM (StEM) algorithm is developed to tackle the computational
problem. Lastly, we apply the proposed method to a social network data
collected from Sina Weibo, and identify the infuential network users as an
interesting application.
| stat |
Unit Level Modeling of Survey Data for Small Area Estimation Under
Informative Sampling: A Comprehensive Overview with Extensions | Model-based small area estimation is frequently used in conjunction with
survey data in order to establish estimates for under-sampled or unsampled
geographies. These models can be specified at either the area-level, or the
unit-level, but unit-level models often offer potential advantages such as more
precise estimates and easy spatial aggregation. Nevertheless, relative to
area-level models, literature on unit-level models is less prevalent. In
modeling small areas at the unit level, challenges often arise as a consequence
of the informative sampling mechanism used to collect the survey data. This
paper provides a comprehensive methodological review for unit-level models
under informative sampling, with an emphasis on Bayesian approaches. To provide
insight into the differences between methods, we conduct a simulation study
that compares several of the described approaches. In addition, the methods
used for simulation are further illustrated through an application to the
American Community Survey. Finally, we present several extensions and areas for
future research.
| stat |
Which practical interventions does the do-operator refer to in causal
inference? Illustration on the example of obesity and cancer | For exposures $X$ like obesity, no precise and unambiguous definition exists
for the hypothetical intervention $do(X = x_0)$. This has raised concerns about
the relevance of causal effects estimated from observational studies for such
exposures. Under the framework of structural causal models, we study how the
effect of $do(X = x_0)$ relates to the effect of interventions on causes of
$X$. We show that for interventions focusing on causes of $X$ that affect the
outcome through $X$ only, the effect of $do(X = x_0)$ equals the effect of the
considered intervention. On the other hand, for interventions on causes $W$ of
$X$ that affect the outcome not only through $X$, we show that the effect of
$do(X = x_0)$ only partly captures the effect of the intervention. In
particular, under simple causal models (e.g., linear models with no
interaction), the effect of $do(X = x_0)$ can be seen as an indirect effect of
the intervention on $W$.
| stat |
Study on estimators of the PDF and CDF of the one parameter polynomial
exponential distribution | In this article, we have considered one parameter polynomial exponential
(OPPE) distribution. The exponential, Lindley, length-biased Lindley and
Sujatha distribution are particular cases. Two estimators viz, MLE and UMVUE of
the PDF and the CDF of the OPPE distribution have been discussed. The
estimation issue of the length-biased Lindley and Sujatha distribution have
been considered in detail. The estimators have been compared in MSE sense.
Monte Carlo simulations and real data analysis are performed to compare the
performances of the proposed methods of estimation.
| stat |
Variational Inference for the Smoothing Distribution in Dynamic Probit
Models | Recently, Fasano, Rebaudo, Durante and Petrone (2019) provided closed-form
expressions for the filtering, predictive and smoothing distributions of
multivariate dynamic probit models, leveraging on unified skew-normal
distribution properties. This allows to develop algorithms to draw independent
and identically distributed samples from such distributions, as well as
sequential Monte Carlo procedures for the filtering and predictive
distributions, allowing to overcome computational bottlenecks that may arise
for large sample sizes. In this paper, we briefly review the above-mentioned
closed-form expressions, mainly focusing on the smoothing distribution of the
univariate dynamic probit. We develop a variational Bayes approach, extending
the partially factorized mean-field variational approximation introduced by
Fasano, Durante and Zanella (2019) for the static binary probit model to the
dynamic setting. Results are shown for a financial application.
| stat |
High-Dimensional Multi-Task Averaging and Application to Kernel Mean
Embedding | We propose an improved estimator for the multi-task averaging problem, whose
goal is the joint estimation of the means of multiple distributions using
separate, independent data sets. The naive approach is to take the empirical
mean of each data set individually, whereas the proposed method exploits
similarities between tasks, without any related information being known in
advance. First, for each data set, similar or neighboring means are determined
from the data by multiple testing. Then each naive estimator is shrunk towards
the local average of its neighbors. We prove theoretically that this approach
provides a reduction in mean squared error. This improvement can be significant
when the dimension of the input space is large, demonstrating a "blessing of
dimensionality" phenomenon. An application of this approach is the estimation
of multiple kernel mean embeddings, which plays an important role in many
modern applications. The theoretical results are verified on artificial and
real world data.
| stat |
What is an Ordinal Latent Trait Model? | Although various polytomous item response models are considered to be ordinal
models there seems no general definition of an ordinal model available.
Alternative concepts of ordinal models are discussed and it is shown that they
coincide for classical unidimensional models. For multidimensional models the
definition of an ordinal model refers to specific traits in the
multidimensional space of traits. The objective is to provide a theoretical
framework for ordinal models. Practical considerations concerning the strength
of the link between the latent trait and the order of categories are considered
briefly.
| stat |
Primary analysis method for incomplete CD4 count data from IMPI trial
and other trials with similar setting | The National Research Council panel on prevention and treatment of missing
data in clinical trials recommends that primary analysis methods are carefully
selected before appropriate sensitivity analysis methods can be chosen. In this
paper, we recommend an appropriate primary analysis method for handling CD4
count data from the IMPI trial and trials with similar settings. The estimand
of interest in the IMPI trial is the effectiveness estimand hypothesis. We
discussed, compared, and contrasted results from complete case analysis and
simple imputation methods, with the direct-likelihood and multiple imputation
methods. The simple imputation methods produced biased estimates of treatment
effect. However, the maximum likelihood and the multiple imputation methods
produced consistent estimates of treatment effect. The maximum likelihood or
the multiple imputation approaches produced unbiased and consistent estimates.
Therefore, either the maximum likelihood or the multiple imputation methods,
under the assumption that the data are missing at random can be considered as
primary analysis method when one is considering sensitivity analysis to dropout
using the CD4 count data from the IMPI trial and other trials with similar
settings.
| stat |
Robust decision analysis under severe uncertainty and ambiguous
tradeoffs: an invasive species case study | Bayesian decision analysis is a useful method for risk management decisions,
but is limited in its ability to consider severe uncertainty in knowledge, and
value ambiguity in management objectives. We study the use of robust Bayesian
decision analysis to handle problems where one or both of these issues arise.
The robust Bayesian approach models severe uncertainty through bounds on
probability distributions, and value ambiguity through bounds on utility
functions. To incorporate data, standard Bayesian updating is applied on the
entire set of distributions. To elicit our expert's utility representing the
value of different management objectives, we use a modified version of the
swing weighting procedure that can cope with severe value ambiguity. We
demonstrate these methods on an environmental management problem to eradicate
an alien invasive marmorkrebs recently discovered in Sweden, which needed a
rapid response despite substantial knowledge gaps if the species was still
present (i.e. severe uncertainty) and the need for difficult tradeoffs and
competing interests (i.e. value ambiguity). We identify that the decision
alternatives to drain the system and remove individuals in combination with
dredging and sieving with or without a degradable biocide, or increasing pH,
are consistently bad under the entire range of probability and utility bounds.
This case study shows how robust Bayesian decision analysis provides a
transparent methodology for integrating information in risk management problems
where little data are available and/or where the tradeoffs ambiguous.
| stat |
Functional Sequential Treatment Allocation with Covariates | We consider a multi-armed bandit problem with covariates. Given a realization
of the covariate vector, instead of targeting the treatment with highest
conditional expectation, the decision maker targets the treatment which
maximizes a general functional of the conditional potential outcome
distribution, e.g., a conditional quantile, trimmed mean, or a socio-economic
functional such as an inequality, welfare or poverty measure. We develop
expected regret lower bounds for this problem, and construct a near minimax
optimal assignment policy.
| stat |
Learning Influence-Receptivity Network Structure with Guarantee | Traditional works on community detection from observations of information
cascade assume that a single adjacency matrix parametrizes all the observed
cascades. However, in reality the connection structure usually does not stay
the same across cascades. For example, different people have different topics
of interest, therefore the connection structure depends on the
information/topic content of the cascade. In this paper we consider the case
where we observe a sequence of noisy adjacency matrices triggered by
information/event with different topic distributions. We propose a novel latent
model using the intuition that a connection is more likely to exist between two
nodes if they are interested in similar topics, which are common with the
information/event. Specifically, we endow each node with two node-topic
vectors: an influence vector that measures how influential/authoritative they
are on each topic; and a receptivity vector that measures how
receptive/susceptible they are to each topic. We show how these two node-topic
structures can be estimated from observed adjacency matrices with theoretical
guarantee on estimation error, in cases where the topic distributions of the
information/event are known, as well as when they are unknown. Experiments on
synthetic and real data demonstrate the effectiveness of our model and superior
performance compared to state-of-the-art methods.
| stat |
Learning Optimal Distributionally Robust Individualized Treatment Rules | Recent development in the data-driven decision science has seen great
advances in individualized decision making. Given data with individual
covariates, treatment assignments and outcomes, policy makers best
individualized treatment rule (ITR) that maximizes the expected outcome, known
as the value function. Many existing methods assume that the training and
testing distributions are the same. However, the estimated optimal ITR may have
poor generalizability when the training and testing distributions are not
identical. In this paper, we consider the problem of finding an optimal ITR
from a restricted ITR class where there is some unknown covariate changes
between the training and testing distributions. We propose a novel
distributionally robust ITR (DR-ITR) framework that maximizes the worst-case
value function across the values under a set of underlying distributions that
are "close" to the training distribution. The resulting DR-ITR can guarantee
the performance among all such distributions reasonably well. We further
propose a calibrating procedure that tunes the DR-ITR adaptively to a small
amount of calibration data from a target population. In this way, the
calibrated DR-ITR can be shown to enjoy better generalizability than the
standard ITR based on our numerical studies.
| stat |
Lasso Meets Horseshoe : A Survey | The goal of this paper is to contrast and survey the major advances in two of
the most commonly used high-dimensional techniques, namely, the Lasso and
horseshoe regularization. Lasso is a gold standard for predictor selection
while horseshoe is a state-of-the-art Bayesian estimator for sparse signals.
Lasso is fast and scalable and uses convex optimization whilst the horseshoe is
non-convex. Our novel perspective focuses on three aspects: (i) theoretical
optimality in high dimensional inference for the Gaussian sparse model and
beyond, (ii) efficiency and scalability of computation and (iii) methodological
development and performance.
| stat |
Optimal Uncertainty-guided Neural Network Training | The neural network (NN)-based direct uncertainty quantification (UQ) methods
have achieved the state of the art performance since the first inauguration,
known as the lower-upper-bound estimation (LUBE) method. However,
currently-available cost functions for uncertainty guided NN training are not
always converging and all converged NNs are not generating optimized prediction
intervals (PIs). Moreover, several groups have proposed different quality
criteria for PIs. These raise a question about their relative effectiveness.
Most of the existing cost functions of uncertainty guided NN training are not
customizable and the convergence of training is uncertain. Therefore, in this
paper, we propose a highly customizable smooth cost function for developing NNs
to construct optimal PIs. The optimized average width of PIs, PI-failure
distances and the PI coverage probability (PICP) are computed for the test
dataset. The performance of the proposed method is examined for the wind power
generation and the electricity demand data. Results show that the proposed
method reduces variation in the quality of PIs, accelerates the training, and
improves convergence probability from 99.2% to 99.8%.
| stat |
Low-Rank Covariance Function Estimation for Multidimensional Functional
Data | Multidimensional function data arise from many fields nowadays. The
covariance function plays an important role in the analysis of such
increasingly common data. In this paper, we propose a novel nonparametric
covariance function estimation approach under the framework of reproducing
kernel Hilbert spaces (RKHS) that can handle both sparse and dense functional
data. We extend multilinear rank structures for (finite-dimensional) tensors to
functions, which allow for flexible modeling of both covariance operators and
marginal structures. The proposed framework can guarantee that the resulting
estimator is automatically semi-positive definite, and can incorporate various
spectral regularizations. The trace-norm regularization in particular can
promote low ranks for both covariance operator and marginal structures. Despite
the lack of a closed form, under mild assumptions, the proposed estimator can
achieve unified theoretical results that hold for any relative magnitudes
between the sample size and the number of observations per sample field, and
the rate of convergence reveals the "phase-transition" phenomenon from sparse
to dense functional data. Based on a new representer theorem, an ADMM algorithm
is developed for the trace-norm regularization. The appealing numerical
performance of the proposed estimator is demonstrated by a simulation study and
the analysis of a dataset from the Argo project.
| stat |
Random clique covers for graphs with local density and global sparsity | Large real-world graphs tend to be sparse, but they often contain many
densely connected subgraphs and exhibit high clustering coefficients. While
recent random graph models can capture this sparsity, they ignore the local
density, or vice versa. We develop a Bayesian nonparametric graph model based
on random edge clique covers, and show that this model can capture power law
degree distribution, global sparsity and non-vanishing local clustering
coefficient. This distribution can be used directly as a prior on observed
graphs, or as part of a hierarchical Bayesian model for inferring latent graph
structures.
| stat |
The Ridgelet Prior: A Covariance Function Approach to Prior
Specification for Bayesian Neural Networks | Bayesian neural networks attempt to combine the strong predictive performance
of neural networks with formal quantification of uncertainty associated with
the predictive output in the Bayesian framework. However, it remains unclear
how to endow the parameters of the network with a prior distribution that is
meaningful when lifted into the output space of the network. A possible
solution is proposed that enables the user to posit an appropriate covariance
function for the task at hand. Our approach constructs a prior distribution for
the parameters of the network, called a ridgelet prior, that approximates the
posited covariance structure in the output space of the network. The approach
is rooted in the ridgelet transform and we establish both finite-sample-size
error bounds and the consistency of the approximation of the covariance
function in a limit where the number of hidden units is increased. Our
experimental assessment is limited to a proof-of-concept, where we demonstrate
that the ridgelet prior can out-perform an unstructured prior on regression
problems for which an informative covariance function can be a priori provided.
| stat |
Small-sample performance and underlying assumptions of a bootstrap-based
inference method for a general analysis of covariance model with possibly
heteroskedastic and nonnormal errors | It is well known that the standard F test is severely affected by
heteroskedasticity in unbalanced analysis of covariance (ANCOVA) models.
Currently available potential remedies for such a scenario are based on
heteroskedasticity-consistent covariance matrix estimation (HCCME). However,
the HCCME approach tends to be liberal in small samples. Therefore, in the
present manuscript, we propose a combination of HCCME and a wild bootstrap
technique, with the aim of improving the small-sample performance. We precisely
state a set of assumptions for the general ANCOVA model and discuss their
practical interpretation in detail, since this issue may have been somewhat
neglected in applied research so far. We prove that these assumptions are
sufficient to ensure the asymptotic validity of the combined HCCME-wild
bootstrap ANCOVA. The results of our simulation study indicate that our
proposed test remedies the problems of the ANCOVA F test and its
heteroskedasticity-consistent alternatives in small to moderate sample size
scenarios. Our test only requires very mild conditions, thus being applicable
in a broad range of real-life settings, as illustrated by the detailed
discussion of a dataset from preclinical research on spinal cord injury. Our
proposed method is ready-to-use and allows for valid hypothesis testing in
frequently encountered settings (e.g., comparing group means while adjusting
for baseline measurements in a randomized controlled clinical trial).
| stat |
Bayesian Conditional Transformation Models | Recent developments in statistical regression methodology establish flexible
relationships between all parameters of the response distribution and the
covariates. This shift away from pure mean regression is just one example and
is further intensified by conditional transformation models (CTMs). They aim to
infer the entire conditional distribution directly by applying a transformation
function that transforms the response conditionally on a set of covariates
towards a simple log-concave reference distribution. Thus, CTMs allow not only
variance, kurtosis and skewness but the complete conditional distribution
function to depend on the explanatory variables. In this article, we propose a
Bayesian notion of conditional transformation models (BCTM) for discrete and
continuous responses in the presence of random censoring. Rather than relying
on simple polynomials, we implement a spline-based parametrization for
monotonic effects that are supplemented with smoothness penalties. Furthermore,
we are able to benefit from the Bayesian paradigm directly via easily
obtainable credible intervals and other quantities without relying on large
sample approximations. A simulation study demonstrates the competitiveness of
our approach against its likelihood-based counterpart, most likely
transformations (MLTs) and Bayesian additive models of location, scale and
shape (BAMLSS). Three applications illustrate the versatility of the BCTMs in
problems involving real world data.
| 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.