title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Penalized regression with multiple loss functions and selection by vote | This article considers a linear model in a high dimensional data scenario. We
propose a process which uses multiple loss functions both to select relevant
predictors and to estimate parameters, and study its asymptotic properties.
Variable selection is conducted by a procedure called "vote", which aggregates
results from penalized loss functions. Using multiple objective functions
separately simplifies algorithms and allows parallel computing, which is
convenient and fast. As a special example we consider a quantile regression
model, which optimally combines multiple quantile levels. We show that the
resulting estimators for the parameter vector are asymptotically efficient.
Simulations and a data application confirm the three main advantages of our
approach: (a) reducing the false discovery rate of variable selection; (b)
improving the quality of parameter estimation; (c) increasing the efficiency of
computation.
| stat |
Randomized Predictive P-values: A Versatile Model Diagnostic Tool with
Unified Reference Distribution | Examining residuals such as Pearson and deviance residuals, is a standard
tool for assessing normal regression. However, for discrete response, these
residuals cluster on lines corresponding to distinct response values. Their
distributions are far from normality; graphical and quantitative inspection of
these residuals provides little information for model diagnosis. Marshall and
Spiegelhalter (2003) defined a cross-validatory predictive p-value for
identifying outliers. Predictive p-values are uniformly distributed for
continuous response but not for discrete response. We propose to use randomized
predictive p-values (RPP) for diagnosing models with discrete responses. RPPs
can be transformed to "residuals" with normal distribution, called NRPPs by us.
NRPPs can be used to diagnose all regression models with scalar response using
the same way for diagnosing normal regression. The NRPPs are nearly the same as
the randomized quantile residuals (RQR), which are previously proposed by Dunn
and Smyth (1996) but remain little known by statisticians. This paper provides
an exposition of RQR using the RPP perspective. The contributions of this
exposition include: (1) we give a rigorous proof of uniformity of RPP and
illustrative examples to explain the uniformity under the true model; (2) we
conduct extensive simulation studies to demonstrate the normality of NRPPs
under the true model; (3) our simulation studies also show that the NRPP method
is a versatile diagnostic tool for detecting many kinds of model inadequacies
due to lack of complexity. The effectiveness of NRPP is further demonstrated
with a health utilization dataset.
| stat |
Regularized Weighted Chebyshev Approximations for Support Estimation | We introduce a new method for estimating the support size of an unknown
distribution which provably matches the performance bounds of the
state-of-the-art techniques in the area and outperforms them in practice. In
particular, we present both theoretical and computer simulation results that
illustrate the utility and performance improvements of our method. The
theoretical analysis relies on introducing a new weighted Chebyshev polynomial
approximation method, jointly optimizing the bias and variance components of
the risk, and combining the weighted minmax polynomial approximation method
with discretized semi-infinite programming solvers. Such a setting allows for
casting the estimation problem as a linear program (LP) with a small number of
variables and constraints that may be solved as efficiently as the original
Chebyshev approximation problem. Our technique is tested on synthetic data and
used to address an important problem in computational biology - estimating the
number of bacterial genera in the human gut. On synthetic datasets, for
practically relevant sample sizes, we observe significant improvements in the
value of the worst-case risk compared to existing methods. For the
bioinformatics application, using metagenomic data from the NIH Human Gut and
the American Gut Microbiome Projects, we generate a list of frequencies of
bacterial taxa that allows us to estimate the number of bacterial genera to
approximately 2300.
| stat |
Default Bayes Factors for Testing the (In)equality of Several Population
Variances | Testing the (in)equality of variances is an important problem in many
statistical applications. We develop default Bayes factor tests to assess the
(in)equality of two or more population variances, as well as a test for whether
the population variance equals a specific value. The resulting test can be used
to check assumptions for commonly used procedures such as the $t$-test or
ANOVA, or test substantive hypotheses concerning variances directly. We further
extend the Bayes factor to allow $\mathcal{H}_0$ to have a null-region.
Researchers may have directed hypotheses such as $\sigma_1^2 > \sigma_2^2$, or
want to combine hypotheses about equality with hypotheses about inequality, for
example $\sigma_1^2 = \sigma_2^2 > (\sigma_3^2, \sigma_4^2)$. We generalize our
Bayes factor to accommodate such hypotheses for $K > 2$ groups. We show that
our Bayes factor fulfills a number of desiderata, provide practical examples
illustrating the method, and compare it to a recently proposed fractional Bayes
factor procedure by B\"oing-Messing & Mulder (2018). Our procedure is
implemented in the R package $bfvartest$.
| stat |
Nonparametric Matrix Response Regression with Application to Brain
Imaging Data Analysis | With the rapid growth of neuroimaging technologies, a great effort has been
dedicated recently to investigate the dynamic changes in brain activity.
Examples include time course calcium imaging and dynamic brain functional
connectivity. In this paper, we propose a novel nonparametric matrix response
regression model to characterize the nonlinear association between 2D image
outcomes and predictors such as time and patient information. Our estimation
procedure can be formulated as a nuclear norm regularization problem, which can
capture the underlying low-rank structure of the dynamic 2D images. We present
a computationally efficient algorithm, derive the asymptotic theory and show
that the method outperforms other existing approaches in simulations. We then
apply the proposed method to a calcium imaging study for estimating the change
of fluorescent intensities of neurons, and an electroencephalography study for
a comparison in the dynamic connectivity covariance matrices between alcoholic
and control individuals. For both studies, the method leads to a substantial
improvement in prediction error.
| stat |
Regularisation of Neural Networks by Enforcing Lipschitz Continuity | We investigate the effect of explicitly enforcing the Lipschitz continuity of
neural networks with respect to their inputs. To this end, we provide a simple
technique for computing an upper bound to the Lipschitz constant---for multiple
$p$-norms---of a feed forward neural network composed of commonly used layer
types. Our technique is then used to formulate training a neural network with a
bounded Lipschitz constant as a constrained optimisation problem that can be
solved using projected stochastic gradient methods. Our evaluation study shows
that the performance of the resulting models exceeds that of models trained
with other common regularisers. We also provide evidence that the
hyperparameters are intuitive to tune, demonstrate how the choice of norm for
computing the Lipschitz constant impacts the resulting model, and show that the
performance gains provided by our method are particularly noticeable when only
a small amount of training data is available.
| stat |
Linear Convergence of Adaptive Stochastic Gradient Descent | We prove that the norm version of the adaptive stochastic gradient method
(AdaGrad-Norm) achieves a linear convergence rate for a subset of either
strongly convex functions or non-convex functions that satisfy the Polyak
Lojasiewicz (PL) inequality. The paper introduces the notion of Restricted
Uniform Inequality of Gradients (RUIG)---which is a measure of the
balanced-ness of the stochastic gradient norms---to depict the landscape of a
function. RUIG plays a key role in proving the robustness of AdaGrad-Norm to
its hyper-parameter tuning in the stochastic setting. On top of RUIG, we
develop a two-stage framework to prove the linear convergence of AdaGrad-Norm
without knowing the parameters of the objective functions. This framework can
likely be extended to other adaptive stepsize algorithms. The numerical
experiments validate the theory and suggest future directions for improvement.
| stat |
Regression Trees for Cumulative Incidence Functions | The use of cumulative incidence functions for characterizing the risk of one
type of event in the presence of others has become increasingly popular over
the past decade. The problems of modeling, estimation and inference have been
treated using parametric, nonparametric and semi-parametric methods. Efforts to
develop suitable extensions of machine learning methods, such as regression
trees and related ensemble methods, have begun only recently. In this paper, we
develop a novel approach to building regression trees for estimating cumulative
incidence curves in a competing risks setting. The proposed methods employ
augmented estimators of the Brier score risk as the primary basis for building
and pruning trees. The proposed methods are easily implemented using the R
statistical software package. Simulation studies demonstrate the utility of our
approach in the competing risks setting. Data from the Radiation Therapy
Oncology Group (trial 9410) is used to illustrate these new methods.
| stat |
The Murphy Decomposition and the Calibration-Resolution Principle: A New
Perspective on Forecast Evaluation | I provide a unifying perspective on forecast evaluation, characterizing
accurate forecasts of all types, from simple point to complete probabilistic
forecasts, in terms of two fundamental underlying properties, autocalibration
and resolution, which can be interpreted as describing a lack of systematic
mistakes and a high information content. This "calibration-resolution
principle" gives a new insight into the nature of forecasting and generalizes
the famous sharpness principle by Gneiting et al. (2007) from probabilistic to
all types of forecasts. It amongst others exposes the shortcomings of several
widely used forecast evaluation methods. The principle is based on a fully
general version of the Murphy decomposition of loss functions, which I provide.
Special cases of this decomposition are well-known and widely used in
meteorology.
Besides using the decomposition in this new theoretical way, after having
introduced it and the underlying properties in a proper theoretical framework,
accompanied by an illustrative example, I also employ it in its classical sense
as a forecast evaluation method as the meteorologists do: As such, it unveils
the driving forces behind forecast errors and complements classical forecast
evaluation methods. I discuss estimation of the decomposition via kernel
regression and then apply it to popular economic forecasts. Analysis of mean
forecasts from the US Survey of Professional Forecasters and quantile forecasts
derived from Bank of England fan charts indeed yield interesting new insights
and highlight the potential of the method.
| stat |
Chaotic Time Series Prediction using Spatio-Temporal RBF Neural Networks | Due to the dynamic nature, chaotic time series are difficult predict. In
conventional signal processing approaches signals are treated either in time or
in space domain only. Spatio-temporal analysis of signal provides more
advantages over conventional uni-dimensional approaches by harnessing the
information from both the temporal and spatial domains. Herein, we propose an
spatio-temporal extension of RBF neural networks for the prediction of chaotic
time series. The proposed algorithm utilizes the concept of time-space
orthogonality and separately deals with the temporal dynamics and spatial
non-linearity(complexity) of the chaotic series. The proposed RBF architecture
is explored for the prediction of Mackey-Glass time series and results are
compared with the standard RBF. The spatio-temporal RBF is shown to out perform
the standard RBFNN by achieving significantly reduced estimation error.
| stat |
Logistic Box-Cox Regression to Assess the Shape and Median Effect under
Uncertainty about Model Specification | The shape of the relationship between a continuous exposure variable and a
binary disease variable is often central to epidemiologic investigations. This
paper investigates a number of issues surrounding inference and the shape of
the relationship. Presuming that the relationship can be expressed in terms of
regression coefficients and a shape parameter, we investigate how well the
shape can be inferred in settings which might typify epidemiologic
investigations and risk assessment. We also consider a suitable definition of
the median effect of exposure, and investigate how precisely this can be
inferred. This is done both in the case of using a model acknowledging
uncertainty about the shape parameter and in the case of ignoring this
uncertainty and using a two-step method, where in step one we transform the
predictor and in step two we fit a simple linear model with transformed
predictor. All these investigations require a family of exposure-disease
relationships indexed by a shape parameter. For this purpose, we employ a
family based on the Box-Cox transformation.
| stat |
Estimating the effectiveness of permanent price reductions for competing
products using multivariate Bayesian structural time series models | The Florence branch of an Italian supermarket chain recently implemented a
strategy that permanently lowered the price of numerous store brands in several
product categories. To quantify the impact of such a policy change, researchers
often use synthetic control methods for estimating causal effects when a subset
of units receive a single persistent treatment, and the rest are unaffected by
the change. In our applications, however, competitor brands not assigned to
treatment are likely impacted by the intervention because of substitution
effects; more broadly, this type of interference occurs whenever the treatment
assignment of one unit affects the outcome of another. This paper extends the
synthetic control methods to accommodate partial interference, allowing
interference within predefined groups but not between them. Focusing on a class
of causal estimands that capture the effect both on the treated and control
units, we develop a multivariate Bayesian structural time series model for
generating synthetic controls that would have occurred in the absence of an
intervention enabling us to estimate our novel effects. In a simulation study,
we explore our Bayesian procedure's empirical properties and show that it
achieves good frequentists coverage even when the model is misspecified. We use
our new methodology to make causal statements about the impact on sales of the
affected store brands and their direct competitors. Our proposed approach is
implemented in the CausalMBSTS R package.
| stat |
Multivariate prediction of mixed, multilevel, sequential outcomes
arising from in vitro fertilisation | In vitro fertilization (IVF) comprises a sequence of interventions concerned
with the creation and culture of embryos which are then transferred to the
patient's uterus. While the clinically important endpoint is birth, the
responses to each stage of treatment contain additional information about the
reasons for success or failure. As such, the ability to predict not only the
overall outcome of the cycle, but also the stage-specific responses, can be
useful. This could be done by developing separate models for each response
variable, but recent work has suggested that it may be advantageous to use a
multivariate approach to model all outcomes simultaneously. Here, joint
analysis of the sequential responses is complicated by mixed outcome types
defined at two levels (patient and embryo). A further consideration is whether
and how to incorporate information about the response at each stage in models
for subsequent stages. We develop a case study using routinely collected data
from a large reproductive medicine unit in order to investigate the feasibility
and potential utility of multivariate prediction in IVF. We consider two
possible scenarios. In the first, stage-specific responses are to be predicted
prior to treatment commencement. In the second, responses are predicted
dynamically, using the outcomes of previous stages as predictors. In both
scenarios, we fail to observe benefits of joint modelling approaches compared
to fitting separate regression models for each response variable.
| stat |
Network Autocorrelation Models with Egocentric Data | Network autocorrelation models have been widely used for decades to model the
joint distribution of the attributes of a network's actors. This class of
models can estimate both the effect of individual characteristics as well as
the network effect, or social influence, on some actor attribute of interest.
Collecting data on the entire network, however, is very often infeasible or
impossible if the network boundary is unknown or difficult to define. Obtaining
egocentric network data overcomes these obstacles, but as of yet there has been
no clear way to model this type of data and still appropriately capture the
network effect on the actor attributes in a way that is compatible with a joint
distribution on the full network data. This paper adapts the class of network
autocorrelation models to handle egocentric data. The proposed methods thus
incorporate the complex dependence structure of the data induced by the network
rather than simply using ad hoc measures of the egos' networks to model the
mean structure, and can estimate the network effect on the actor attribute of
interest. The vast quantities of unknown information about the network can be
succinctly represented in such a way that only depends on the number of alters
in the egocentric network data and not on the total number of actors in the
network. Estimation is done within a Bayesian framework. A simulation study is
performed to evaluate the estimation performance, and an egocentric data set is
analyzed where the aim is to determine if there is a network effect on
environmental mastery, an important aspect of psychological well-being.
| stat |
Flexible Bayesian modelling in dichotomous item response theory using
mixtures of skewed item curves | Most Item Response Theory (IRT) models for dichotomous responses are based on
probit or logit link functions which assume a symmetric relationship between
the responses and the latent traits of individuals submitted to a test. Such an
assumption restricts the use of such models to situations in which all items
have symmetric behavior. Similar constraint is imposed by the asymmetric models
proposed in the literature as it is required that all items have an asymmetric
behavior. Such assumptions are inappropriate for great part of the tests which,
in general, are composed by both symmetric and asymmetric items. Furthermore, a
straightforward extension of the existing models in the literature of would
require a prior selection of the items' symmetry/asymmetry status. This paper
proposes a Bayesian IRT model that accounts for symmetric and asymmetric items
in a flexible though parsimonious way. That is achieved by assigning a
point-mass mixture prior to the skewness parameter of the item, allowing for an
analysis under the model selection or model averaging approaches. Asymmetric
item curves are design through the centred skew normal distribution which has a
particularly appealing parametrisation in terms of parameter interpretation and
computational efficiency. An efficient MCMC algorithm is proposed to perform
Bayesian inference and its performance is investigated in some simulated
examples. Finally, the proposed methodology is applied to a data set from a
large scale educational exam in Brazil.
| stat |
Housing Search in the Age of Big Data: Smarter Cities or the Same Old
Blind Spots? | Housing scholars stress the importance of the information environment in
shaping housing search behavior and outcomes. Rental listings have increasingly
moved online over the past two decades and, in turn, online platforms like
Craigslist are now central to the search process. Do these technology platforms
serve as information equalizers or do they reflect traditional information
inequalities that correlate with neighborhood sociodemographics? We synthesize
and extend analyses of millions of US Craigslist rental listings and find they
supply significantly different volumes, quality, and types of information in
different communities. Technology platforms have the potential to broaden,
diversify, and equalize housing search information, but they rely on landlord
behavior and, in turn, likely will not reach this potential without a
significant redesign or policy intervention. Smart cities advocates hoping to
build better cities through technology must critically interrogate technology
platforms and big data for systematic biases.
| stat |
A liberal type I error rate for studies in precision medicine | We introduce a new multiple type I error criterion for clinical trials with
multiple populations. Such trials are of interest in precision medicine where
the goal is to develop treatments that are targeted to specific sub-populations
defined by genetic and/or clinical biomarkers. The new criterion is based on
the observation that not all type I errors are relevant to all patients in the
overall population. If disjoint sub-populations are considered, no multiplicity
adjustment appears necessary, since a claim in one sub-population does not
affect patients in the other ones. For intersecting sub-populations we suggest
to control the average multiple type error rate, i.e. the probably that a
randomly selected patient will be exposed to an inefficient treatment. We call
this the population-wise error rate, exemplify it by a number of examples and
illustrate how to control it with an adjustment of critical boundaries or
adjusted p-values. We furthermore define corresponding simultaneous confidence
intervals. We finally illustrate the power gain achieved by passing from
family-wise to population-wise error rate control with two simple examples and
a recently suggest multiple testing approach for umbrella trials.
| stat |
Protection Against Reconstruction and Its Applications in Private
Federated Learning | In large-scale statistical learning, data collection and model fitting are
moving increasingly toward peripheral devices---phones, watches, fitness
trackers---away from centralized data collection. Concomitant with this rise in
decentralized data are increasing challenges of maintaining privacy while
allowing enough information to fit accurate, useful statistical models. This
motivates local notions of privacy---most significantly, local differential
privacy, which provides strong protections against sensitive data
disclosures---where data is obfuscated before a statistician or learner can
even observe it, providing strong protections to individuals' data. Yet local
privacy as traditionally employed may prove too stringent for practical use,
especially in modern high-dimensional statistical and machine learning
problems. Consequently, we revisit the types of disclosures and adversaries
against which we provide protections, considering adversaries with limited
prior information and ensuring that with high probability, ensuring they cannot
reconstruct an individual's data within useful tolerances. By reconceptualizing
these protections, we allow more useful data release---large privacy parameters
in local differential privacy---and we design new (minimax) optimal locally
differentially private mechanisms for statistical learning problems for
\emph{all} privacy levels. We thus present practicable approaches to
large-scale locally private model training that were previously impossible,
showing theoretically and empirically that we can fit large-scale image
classification and language models with little degradation in utility.
| stat |
The consequences of checking for zero-inflation and overdispersion in
the analysis of count data | Count data are ubiquitous in ecology and the Poisson generalized linear model
(GLM) is commonly used to model the association between counts and explanatory
variables of interest. When fitting this model to the data, one typically
proceeds by first confirming that the data is not overdispersed and that there
is no excess of zeros. If the data appear to be overdispersed or if there is
any zero-inflation, key assumptions of the Poison GLM may be violated and
researchers will then typically consider alternatives to the Poison GLM. An
important question is whether the potential model selection bias introduced by
this data-driven multi-stage procedure merits concern. In this paper, we
conduct a large-scale simulation study to investigate the potential
consequences of model selection bias that can arise in the simple scenario of
analyzing a sample of potentially overdispersed, potentially zero-heavy, count
data.
| stat |
Estimating mixed-memberships using the Symmetric Laplacian Inverse
Matrix | Community detection has been well studied in network analysis, and one
popular technique is spectral clustering which is fast and statistically
analyzable for detect-ing clusters for given networks. But the more realistic
case of mixed membership community detection remains a challenge. In this
paper, we propose a new spectral clustering method Mixed-SLIM for mixed
membership community detection. Mixed-SLIM is designed based on the symmetrized
Laplacian inverse matrix (SLIM) (Jing et al. 2021) under the degree-corrected
mixed membership (DCMM) model. We show that this algorithm and its regularized
version Mixed-SLIM {\tau} are asymptotically consistent under mild conditions.
Meanwhile, we provide Mixed-SLIM appro and its regularized version Mixed-SLIM
{\tau}appro by approximating the SLIM matrix when dealing with large networks
in practice. These four Mixed-SLIM methods outperform state-of-art methods in
simulations and substantial empirical datasets for both community detection and
mixed membership community detection problems.
| stat |
SOAR: Simultaneous Or of And Rules for Classification of Positive &
Negative Classes | Algorithmic decision making has proliferated and now impacts our daily lives
in both mundane and consequential ways. Machine learning practitioners make use
of a myriad of algorithms for predictive models in applications as diverse as
movie recommendations, medical diagnoses, and parole recommendations without
delving into the reasons driving specific predictive decisions. Machine
learning algorithms in such applications are often chosen for their superior
performance, however popular choices such as random forest and deep neural
networks fail to provide an interpretable understanding of the predictive
model. In recent years, rule-based algorithms have been used to address this
issue. Wang et al. (2017) presented an or-of-and (disjunctive normal form)
based classification technique that allows for classification rule mining of a
single class in a binary classification; this method is also shown to perform
comparably to other modern algorithms. In this work, we extend this idea to
provide classification rules for both classes simultaneously. That is, we
provide a distinct set of rules for both positive and negative classes. In
describing this approach, we also present a novel and complete taxonomy of
classifications that clearly capture and quantify the inherent ambiguity in
noisy binary classifications in the real world. We show that this approach
leads to a more granular formulation of the likelihood model and a
simulated-annealing based optimization achieves classification performance
competitive with comparable techniques. We apply our method to synthetic as
well as real world data sets to compare with other related methods that
demonstrate the utility of our proposal.
| stat |
Notes on the interpretation of dependence measures | Besides the classical distinction of correlation and dependence, many
dependence measures bear further pitfalls in their application and
interpretation. The aim of this paper is to raise and recall awareness of some
of these limitations by explicitly discussing Pearson's correlation and the
multivariate dependence measures: distance correlation, distance
multicorrelations and their copula versions. The discussed aspects include
types of dependence, bias of empirical measures, influence of marginal
distributions and dimensions.
In general it is recommended to use a proper dependence measure instead of
Pearson's correlation. Moreover, a measure which is distribution-free (at least
in some sense) can help to avoid certain systematic errors. Nevertheless, in a
truly multivariate setting only the p-values of the corresponding independence
tests provide always values with indubitable interpretation.
| stat |
TRAMP: Compositional Inference with TRee Approximate Message Passing | We introduce tramp, standing for TRee Approximate Message Passing, a python
package for compositional inference in high-dimensional tree-structured models.
The package provides an unifying framework to study several approximate message
passing algorithms previously derived for a variety of machine learning tasks
such as generalized linear models, inference in multi-layer networks, matrix
factorization, and reconstruction using non-separable penalties. For some
models, the asymptotic performance of the algorithm can be theoretically
predicted by the state evolution, and the measurements entropy estimated by the
free entropy formalism. The implementation is modular by design: each module,
which implements a factor, can be composed at will with other modules to solve
complex inference tasks. The user only needs to declare the factor graph of the
model: the inference algorithm, state evolution and entropy estimation are
fully automated.
| stat |
Blinded continuous information monitoring of recurrent event endpoints
with time trends in clinical trials | Blinded sample size re-estimation and information monitoring based on blinded
data has been suggested to mitigate risks due to planning uncertainties
regarding nuisance parameters. Motivated by a randomized controlled trial in
pediatric multiple sclerosis (MS), a continuous monitoring procedure for
overdispersed count data was proposed recently. However, this procedure assumed
constant event rates, an assumption often not met in practice. Here we extend
the procedure to accommodate time trends in the event rates considering two
blinded approaches: (a) the mixture approach modeling the number of events by a
mixture of two negative binomial distributions, and (b) the lumping approach
approximating the marginal distribution of the event counts by a negative
binomial distribution. Through simulations the operating characteristics of the
proposed procedures are investigated under decreasing event rates. We find that
the type I error rate is not inflated relevantly by either of the monitoring
procedures, with the exception of strong time dependencies where the procedure
assuming constant rates exhibits some inflation. Furthermore, the procedure
accommodating time trends has generally favorable power properties compared to
the procedure based on constant rates which stops often too late. The proposed
method is illustrated by the clinical trial in pediatric MS.
| stat |
Simulating bridges using confluent diffusions | Diffusions are a fundamental class of models in many fields, including
finance, engineering, and biology. However, their simulation is challenging as
their sample paths are infinite-dimensional and their transition function is
typically intractable. In many statistical settings (such as parameter
inference for discretely observed diffusions), we require simulation techniques
for diffusions conditioned on hitting an endpoint, which introduces further
complication. In this paper we introduce a Markov chain Monte Carlo algorithm
for simulating diffusion bridges which is both exact (in the sense that there
is no discretisation error) and has computational cost that is linear in the
duration of the bridge. Our approach works directly on diffusion path space, by
constructing a proposal (which we term a confluence) with finite computational
cost which is then accepted/rejected using a computable probability. The
novelty of our method is that it only requires the simulation of unconditioned
diffusion sample paths. Our methodology develops a number of simulation
techniques we believe to be of independent interest, related to the crossing
and first passage times of multiple Brownian motions, meanders and bridges, and
Bessel bridges. We apply our methodology to the simulation of Langevin
diffusion bridges, a practical problem arising naturally in statistical fusion
settings.
| stat |
On nearly assumption-free tests of nominal confidence interval coverage
for causal parameters estimated by machine learning | For many causal effect parameters of interest, doubly robust machine learning
(DRML) estimators $\hat{\psi}_{1}$ are the state-of-the-art, incorporating the
good prediction performance of machine learning; the decreased bias of doubly
robust estimators; and the analytic tractability and bias reduction of sample
splitting with cross fitting. Nonetheless, even in the absence of confounding
by unmeasured factors, the nominal $(1 - \alpha)$ Wald confidence interval
$\hat{\psi}_{1} \pm z_{\alpha / 2} \widehat{\mathsf{se}} [\hat{\psi}_{1}]$ may
still undercover even in large samples, because the bias of $\hat{\psi}_{1}$
may be of the same or even larger order than its standard error of order
$n^{-1/2}$.
In this paper, we introduce essentially assumption-free tests that (i) can
falsify the null hypothesis that the bias of $\hat{\psi}_{1}$ is of smaller
order than its standard error, (ii) can provide an upper confidence bound on
the true coverage of the Wald interval, and (iii) are valid under the null
under no smoothness/sparsity assumptions on the nuisance parameters. The tests,
which we refer to as \underline{A}ssumption \underline{F}ree
\underline{E}mpirical \underline{C}overage \underline{T}ests (AFECTs), are
based on a U-statistic that estimates part of the bias of $\hat{\psi}_{1}$.
| stat |
Energy Balancing of Covariate Distributions | Bias in causal comparisons has a direct correspondence with distributional
imbalance of covariates between treatment groups. Weighting strategies such as
inverse propensity score weighting attempt to mitigate bias by either modeling
the treatment assignment mechanism or balancing specified covariate moments.
This paper introduces a new weighting method, called energy balancing, which
instead aims to balance weighted covariate distributions. By directly targeting
distributional imbalance, the proposed weighting strategy can be flexibly
utilized in a wide variety of causal analyses, including the estimation of
average treatment effects and individualized treatment rules. Our energy
balancing weights (EBW) approach has several advantages over existing weighting
techniques. First, it offers a model-free and robust approach for obtaining
covariate balance that does not require tuning parameters, obviating the need
for modeling decisions of secondary nature to the scientific question at hand.
Second, since this approach is based on a genuine measure of distributional
balance, it provides a means for assessing the balance induced by a given set
of weights for a given dataset. Finally, the proposed method is computationally
efficient and has desirable theoretical guarantees under mild conditions. We
demonstrate the effectiveness of this EBW approach in a suite of simulation
experiments, and in studies on the safety of right heart catheterization and
the effect of indwelling arterial catheters.
| stat |
Global estimation of unintended pregnancy and abortion using a Bayesian
hierarchical random walk model | Unintended pregnancy and abortion estimates are needed to inform and motivate
investment in global health programmes and policies. Variability in the
availability and reliability of data poses challenges for producing estimates.
We developed a Bayesian model that simultaneously estimates incidence of
unintended pregnancy and abortion for 195 countries and territories. Our
modelling strategy was informed by the proximate determinants of fertility with
(i) incidence of unintended pregnancy defined by the number of women (grouped
by marital and contraceptive use status) and their respective pregnancy rates,
and (ii) abortion incidence defined by group-specific pregnancies and
propensities to have an abortion. Hierarchical random walk models are used to
estimate country-group-period-specific pregnancy rates and propensities to
abort.
| stat |
On the Use of Information Criteria for Subset Selection in Least Squares
Regression | Least squares (LS)-based subset selection methods are popular in linear
regression modeling. Best subset selection (BS) is known to be NP-hard and has
a computational cost that grows exponentially with the number of predictors.
Recently, Bertsimas (2016) formulated BS as a mixed integer optimization (MIO)
problem and largely reduced the computation overhead by using a well-developed
optimization solver, but the current methodology is not scalable to very large
datasets. In this paper, we propose a novel LS-based method, the best
orthogonalized subset selection (BOSS) method, which performs BS upon an
orthogonalized basis of ordered predictors and scales easily to large problem
sizes. Another challenge in applying LS-based methods in practice is the
selection rule to choose the optimal subset size k. Cross-validation (CV)
requires fitting a procedure multiple times, and results in a selected k that
is random across repeated application to the same dataset. Compared to CV,
information criteria only require fitting a procedure once, but they require
knowledge of the effective degrees of freedom for the fitting procedure, which
is generally not available analytically for complex methods. Since BOSS uses
orthogonalized predictors, we first explore a connection for orthogonal
non-random predictors between BS and its Lagrangian formulation (i.e.,
minimization of the residual sum of squares plus the product of a
regularization parameter and k), and based on this connection propose a
heuristic degrees of freedom (hdf) for BOSS that can be estimated via an
analytically-based expression. We show in both simulations and real data
analysis that BOSS using a proposed Kullback-Leibler based information
criterion AICc-hdf has the strongest performance of all of the LS-based methods
considered and is competitive with regularization methods, with the
computational effort of a single ordinary LS fit.
| stat |
Bayesian Multivariate Spatial Models for Lattice Data with INLA | The INLAMSM package for the R programming language provides a collection of
multivariate spatial models for lattice data that can be used with package INLA
for Bayesian inference. The multivariate spatial models include different
structures to model the spatial variation of the variables and the
between-variables variability. In this way, fitting multivariate spatial models
becomes faster and easier. The use of the different models included in the
package is illustrated using two different datasets: the well-known North
Carolina SIDS data and mortality by three causes of death in Comunidad
Valenciana (Spain).
| stat |
A Feature Weighted Mixed Naive Bayes Model for Monitoring Anomalies in
the Fan System of a Thermal Power Plant | With the increasing intelligence and integration, a great number of
two-valued variables (generally stored in the form of 0 or 1 value) often exist
in large-scale industrial processes. However, these variables cannot be
effectively handled by traditional monitoring methods such as LDA, PCA and PLS.
Recently, a mixed hidden naive Bayesian model (MHNBM) is developed for the
first time to utilize both two-valued and continuous variables for abnormality
monitoring. Although MHNBM is effective, it still has some shortcomings that
need to be improved. For MHNBM, the variables with greater correlation to other
variables have greater weights, which cannot guarantee greater weights are
assigned to the more discriminating variables. In addition, the conditional
probability must be computed based on the historical data. When the training
data is scarce, the conditional probability between continuous variables tends
to be uniformly distributed, which affects the performance of MHNBM. Here a
novel feature weighted mixed naive Bayes model (FWMNBM) is developed to
overcome the above shortcomings. For FWMNBM, the variables that are more
correlated to the class have greater weights, which makes the more
discriminating variables contribute more to the model. At the same time, FWMNBM
does not have to calculate the conditional probability between variables, thus
it is less restricted by the number of training data samples. Compared with
MHNBM, FWMNBM has better performance, and its effectiveness is validated
through the numerical cases of a simulation example and a practical case of
Zhoushan thermal power plant (ZTPP), China.
| stat |
A competing risks interpretation of Hawkes processes | We give a construction of the Hawkes process as a piecewise competing risks
model. We argue that the most natural interpretation of the self-excitation
kernel is the hazard function of a defective random variable. This establishes
a link between desired qualitative features of the process and a parametric
form for the kernel, which we illustrate using examples from the literature.
Two families of cure rate models taken from the survival analysis literature
are proposed as new models for the self-excitation kernel. Finally, we show
that the competing risks viewpoint leads to a general simulation algorithm
which avoids inverting the compensator of the point process or performing an
accept-reject step, and is therefore fast and quite general.
| stat |
Mixed pooling of seasonality for time series forecasting: An application
to pallet transport data | Multiple seasonal patterns play a key role in time series forecasting,
especially for business time series where seasonal effects are often dramatic.
Previous approaches including Fourier decomposition, exponential smoothing, and
seasonal autoregressive integrated moving average (SARIMA) models do not
reflect the distinct characteristics of each period in seasonal patterns. We
propose a mixed hierarchical seasonality (MHS) model. Intermediate parameters
for each seasonal period are first estimated, and a mixture of intermediate
parameters is taken. This results in a model that automatically learns the
relative importance of each seasonality and addresses the interactions between
them. The model is implemented with Stan, a probabilistic language, and was
compared with three existing models on a real-world dataset of pallet transport
from a logistic network. Our new model achieved considerable improvements in
terms of out of sample prediction error (MAPE) and predictive density (ELPD)
compared to complete pooling, Fourier decomposition, and SARIMA model.
| stat |
Penalized model-based clustering of fMRI data | Functional magnetic resonance imaging (fMRI) data have become increasingly
available and are useful for describing functional connectivity (FC), the
relatedness of neuronal activity in regions of the brain. This FC of the brain
provides insight into certain neurodegenerative diseases and psychiatric
disorders, and thus is of clinical importance. To help inform physicians
regarding patient diagnoses, unsupervised clustering of subjects based on FC is
desired, allowing the data to inform us of groupings of patients based on
shared features of connectivity. Since heterogeneity in FC is present even
between patients within the same group, it is important to allow subject-level
differences in connectivity, while still pooling information across patients
within each group to describe group-level FC. To this end, we propose a random
covariance clustering model (RCCM) to concurrently cluster subjects based on
their FC networks, estimate the unique FC networks of each subject, and to
infer shared network features. Although current methods exist for estimating FC
or clustering subjects using fMRI data, our novel contribution is to cluster or
group subjects based on similar FC of the brain while simultaneously providing
group- and subject-level FC network estimates. The competitive performance of
RCCM relative to other methods is demonstrated through simulations in various
settings, achieving both improved clustering of subjects and estimation of FC
networks. Utility of the proposed method is demonstrated with application to a
resting-state fMRI data set collected on 43 healthy controls and 61
participants diagnosed with schizophrenia.
| stat |
Space-Time Covid-19 Bayesian SIR modeling in South Carolina | The Covid-19 pandemic has spread across the world since the beginning of
2020. Many regions have experienced its effects. The state of South Carolina in
the USA has seen cases since early March 2020 and a primary peak in early April
2020. A lockdown was imposed on April 6th but lifting of restrictions started
on April 24th. The daily case and death data as reported by NCHS (deaths) via
the New York Times GitHUB repository have been analyzed and approaches to
modeling of the data are presented. Prediction is also considered and the role
of asymptomatic transmission is assessed as a latent unobserved effect. Two
different time periods are examined and one step prediction is provided.
| stat |
Parameter estimation in models generated by SDE's with symmetric alpha
stable noise | The article considers vector parameter estimators in statistical models
generated by Levy processes. An improved one step estimator is presented that
can be used for improving any other estimator. Combined numerical methods for
optimization problems are proposed. A software has been developed and a
correspondent testing and comparison have been presented.
| stat |
Variance partitioning in multilevel models for count data | A first step when fitting multilevel models to continuous responses is to
explore the degree of clustering in the data. Researchers fit
variance-component models and then report the proportion of variation in the
response that is due to systematic differences between clusters. Equally they
report the response correlation between units within a cluster. These
statistics are popularly referred to as variance partition coefficients (VPCs)
and intraclass correlation coefficients (ICCs). When fitting multilevel models
to categorical (binary, ordinal, or nominal) and count responses, these
statistics prove more challenging to calculate. For categorical response
models, researchers appeal to their latent response formulations and report
VPCs/ICCs in terms of latent continuous responses envisaged to underly the
observed categorical responses. For standard count response models, however,
there are no corresponding latent response formulations. More generally, there
is a paucity of guidance on how to partition the variation. As a result,
applied researchers are likely to avoid or inadequately report and discuss the
substantive importance of clustering and cluster effects in their studies. A
recent article drew attention to a little-known exact algebraic expression for
the VPC/ICC for the special case of the two-level random-intercept Poisson
model. In this article, we make a substantial new contribution. First, we
derive exact VPC/ICC expressions for more flexible negative binomial models
that allows for overdispersion, a phenomenon which often occurs in practice.
Then we derive exact VPC/ICC expressions for three-level and random-coefficient
extensions to these models. We illustrate our work with an application to
student absenteeism.
| stat |
Network Dependence Testing via Diffusion Maps and Distance-Based
Correlations | Deciphering the associations between network connectivity and nodal
attributes is one of the core problems in network science. The dependency
structure and high-dimensionality of networks pose unique challenges to
traditional dependency tests in terms of theoretical guarantees and empirical
performance. We propose an approach to test network dependence via diffusion
maps and distance-based correlations. We prove that the new method yields a
consistent test statistic under mild distributional assumptions on the graph
structure, and demonstrate that it is able to efficiently identify the most
informative graph embedding with respect to the diffusion time. The methodology
is illustrated on both simulated and real data.
| stat |
Maximal couplings of the Metropolis-Hastings algorithm | Couplings play a central role in the analysis of Markov chain Monte Carlo
algorithms and appear increasingly often in the algorithms themselves, e.g. in
convergence diagnostics, parallelization, and variance reduction techniques.
Existing couplings of the Metropolis-Hastings algorithm handle the proposal and
acceptance steps separately and fall short of the upper bound on one-step
meeting probabilities given by the coupling inequality. This paper introduces
maximal couplings which achieve this bound while retaining the practical
advantages of current methods. We consider the properties of these couplings
and examine their behavior on a selection of numerical examples.
| stat |
Neural Density Estimation and Likelihood-free Inference | I consider two problems in machine learning and statistics: the problem of
estimating the joint probability density of a collection of random variables,
known as density estimation, and the problem of inferring model parameters when
their likelihood is intractable, known as likelihood-free inference. The
contribution of the thesis is a set of new methods for addressing these
problems that are based on recent advances in neural networks and deep
learning.
| stat |
Classification Under Human Assistance | Most supervised learning models are trained for full automation. However,
their predictions are sometimes worse than those by human experts on some
specific instances. Motivated by this empirical observation, our goal is to
design classifiers that are optimized to operate under different automation
levels. More specifically, we focus on convex margin-based classifiers and
first show that the problem is NP-hard. Then, we further show that, for support
vector machines, the corresponding objective function can be expressed as the
difference of two functions f = g - c, where g is monotone, non-negative and
{\gamma}-weakly submodular, and c is non-negative and modular. This
representation allows a recently introduced deterministic greedy algorithm, as
well as a more efficient randomized variant of the algorithm, to enjoy
approximation guarantees at solving the problem. Experiments on synthetic and
real-world data from several applications in medical diagnosis illustrate our
theoretical findings and demonstrate that, under human assistance, supervised
learning models trained to operate under different automation levels can
outperform those trained for full automation as well as humans operating alone.
| stat |
Two-sided confidence interval of a binomial proportion: how to choose? | Introduction: estimation of confidence intervals (CIs) of binomial
proportions has been reviewed more than once but the directional
interpretation, distinguishing the overestimation from the underestimation, was
neglected while the sample size and theoretical proportion variances from
experiment to experiment have not been formally taken in account. Herein, we
define and apply new evaluation criteria, then give recommendations for the
practical use of these CIs.
Materials & methods: Google Scholar was used for bibliographic research.
Evaluation criteria were (i) one-sided conditional errors, (ii) one-sided local
average errors assuming a random theoretical proportion and (iii) expected
half-widths of CIs.
Results: Wald's CI did not control any of the risks, even when the expected
number of successes reached 32. The likelihood ratio CI had a better balance
than the logistic Wald CI. The Clopper-Pearson mid-P CI controlled well
one-sided local average errors whereas the simple Clopper-Pearson CI was
strictly conservative on both one-sided conditional errors. The percentile and
basic bootstrap CIs had the same bias order as Wald's CI whereas the
studentized CIs and BCa, modified for discrete bootstrap distributions, were
less biased but not as efficient as the parametric methods. The half-widths of
CIs mirrored local average errors.
Conclusion: we recommend using the Clopper-Pearson mid-P CI for the
estimation of a proportion except for observed-theoretical proportion
comparison under controlled experimental conditions in which the
Clopper-Pearson CI may be better.
| stat |
High-dimensional Multivariate Geostatistics: A Bayesian Matrix-Normal
Approach | Joint modeling of spatially-oriented dependent variables is commonplace in
the environmental sciences, where scientists seek to estimate the relationships
among a set of environmental outcomes accounting for dependence among these
outcomes and the spatial dependence for each outcome. Such modeling is now
sought for massive data sets with variables measured at a very large number of
locations. Bayesian inference, while attractive for accommodating uncertainties
through hierarchical structures, can become computationally onerous for
modeling massive spatial data sets because of its reliance on iterative
estimation algorithms. This manuscript develops a conjugate Bayesian framework
for analyzing multivariate spatial data using analytically tractable posterior
distributions that obviate iterative algorithms. We discuss differences between
modeling the multivariate response itself as a spatial process and that of
modeling a latent process in a hierarchical model. We illustrate the
computational and inferential benefits of these models using simulation studies
and analysis of a Vegetation Index data set with spatially dependent
observations numbering in the millions.
| stat |
Bayesian Variable Selection for Single Index Logistic Model | In the era of big data, variable selection is a key technology for handling
high-dimensional problems with a small sample size but a large number of
covariables. Different variable selection methods were proposed for different
models, such as linear model, logistic model and generalized linear model.
However, fewer works focused on variable selection for single index models,
especially, for single index logistic model, due to the difficulty arose from
the unknown link function and the slow mixing rate of MCMC algorithm for
traditional logistic model. In this paper, we proposed a Bayesian variable
selection procedure for single index logistic model by taking the advantage of
Gaussian process and data augmentation. Numerical results from simulations and
real data analysis show the advantage of our method over the state of arts.
| stat |
A Bayesian approach for clustering skewed data using mixtures of
multivariate normal-inverse Gaussian distributions | Non-Gaussian mixture models are gaining increasing attention for mixture
model-based clustering particularly when dealing with data that exhibit
features such as skewness and heavy tails. Here, such a mixture distribution is
presented, based on the multivariate normal inverse Gaussian (MNIG)
distribution. For parameter estimation of the mixture, a Bayesian approach via
Gibbs sampler is used; for this, a novel approach to simulate univariate
generalized inverse Gaussian random variables and matrix generalized inverse
Gaussian random matrices is provided. The proposed algorithm will be applied to
both simulated and real data. Through simulation studies and real data
analysis, we show parameter recovery and that our approach provides competitive
clustering results compared to other clustering approaches.
| stat |
Diffusion State Distances: Multitemporal Analysis, Fast Algorithms, and
Applications to Biological Networks | Data-dependent metrics are powerful tools for learning the underlying
structure of high-dimensional data. This article develops and analyzes a
data-dependent metric known as diffusion state distance (DSD), which compares
points using a data-driven diffusion process. Unlike related diffusion methods,
DSDs incorporate information across time scales, which allows for the intrinsic
data structure to be inferred in a parameter-free manner. This article develops
a theory for DSD based on the multitemporal emergence of mesoscopic equilibria
in the underlying diffusion process. New algorithms for denoising and dimension
reduction with DSD are also proposed and analyzed. These approaches are based
on a weighted spectral decomposition of the underlying diffusion process, and
experiments on synthetic datasets and real biological networks illustrate the
efficacy of the proposed algorithms in terms of both speed and accuracy.
Throughout, comparisons with related methods are made, in order to illustrate
the distinct advantages of DSD for datasets exhibiting multiscale structure.
| stat |
Implicit Regularization for Optimal Sparse Recovery | We investigate implicit regularization schemes for gradient descent methods
applied to unpenalized least squares regression to solve the problem of
reconstructing a sparse signal from an underdetermined system of linear
measurements under the restricted isometry assumption. For a given
parametrization yielding a non-convex optimization problem, we show that
prescribed choices of initialization, step size and stopping time yield a
statistically and computationally optimal algorithm that achieves the minimax
rate with the same cost required to read the data up to poly-logarithmic
factors. Beyond minimax optimality, we show that our algorithm adapts to
instance difficulty and yields a dimension-independent rate when the
signal-to-noise ratio is high enough. Key to the computational efficiency of
our method is an increasing step size scheme that adapts to refined estimates
of the true solution. We validate our findings with numerical experiments and
compare our algorithm against explicit $\ell_{1}$ penalization. Going from hard
instances to easy ones, our algorithm is seen to undergo a phase transition,
eventually matching least squares with an oracle knowledge of the true support.
| stat |
Markov-switching State Space Models for Uncovering Musical
Interpretation | For concertgoers, musical interpretation is the most important factor in
determining whether or not we enjoy a classical performance. Every performance
includes mistakes---intonation issues, a lost note, an unpleasant sound---but
these are all easily forgotten (or unnoticed) when a performer engages her
audience, imbuing a piece with novel emotional content beyond the vague
instructions inscribed on the printed page. While music teachers use imagery or
heuristic guidelines to motivate interpretive decisions, combining these vague
instructions to create a convincing performance remains the domain of the
performer, subject to the whims of the moment, technical fluency, and taste. In
this research, we use data from the CHARM Mazurka Project---forty-six
professional recordings of Chopin's Mazurka Op. 63 No. 3 by consumate
artists---with the goal of elucidating musically interpretable performance
decisions. Using information on the inter-onset intervals of the note attacks
in the recordings, we apply functional data analysis techniques enriched with
prior information gained from music theory to discover relevant features and
perform hierarchical clustering. The resulting clusters suggest methods for
informing music instruction, discovering listening preferences, and analyzing
performances.
| stat |
Parsimonious Bayesian deep networks | Combining Bayesian nonparametrics and a forward model selection strategy, we
construct parsimonious Bayesian deep networks (PBDNs) that infer
capacity-regularized network architectures from the data and require neither
cross-validation nor fine-tuning when training the model. One of the two
essential components of a PBDN is the development of a special infinite-wide
single-hidden-layer neural network, whose number of active hidden units can be
inferred from the data. The other one is the construction of a greedy
layer-wise learning algorithm that uses a forward model selection criterion to
determine when to stop adding another hidden layer. We develop both Gibbs
sampling and stochastic gradient descent based maximum a posteriori inference
for PBDNs, providing state-of-the-art classification accuracy and interpretable
data subtypes near the decision boundaries, while maintaining low computational
complexity for out-of-sample prediction.
| stat |
Scaleable input gradient regularization for adversarial robustness | In this work we revisit gradient regularization for adversarial robustness
with some new ingredients. First, we derive new per-image theoretical
robustness bounds based on local gradient information. These bounds strongly
motivate input gradient regularization. Second, we implement a scaleable
version of input gradient regularization which avoids double backpropagation:
adversarially robust ImageNet models are trained in 33 hours on four consumer
grade GPUs. Finally, we show experimentally and through theoretical
certification that input gradient regularization is competitive with
adversarial training. Moreover we demonstrate that gradient regularization does
not lead to gradient obfuscation or gradient masking.
| stat |
Uncertain Neighbors: Bayesian Propensity Score Matching For Causal
Inference | We compare the performance of standard nearest-neighbor propensity score
matching with that of an analogous Bayesian propensity score matching
procedure. We show that the Bayesian approach makes better use of available
information, as it makes less arbitrary decisions about which observations to
drop and which ones to keep in the matched sample. We conduct a simulation
study to evaluate the performance of standard and Bayesian nearest-neighbor
matching when matching is done with and without replacement. We then use both
methods to replicate a recent study about the impact of land reform on
guerrilla activity in Colombia.
| stat |
An Efficient $k$-modes Algorithm for Clustering Categorical Datasets | Mining clusters from datasets is an important endeavor in many applications.
The $k$-means algorithm is a popular and efficient, distribution-free approach
for clustering numerical-valued data, but does not apply for categorical-valued
observations. The $k$-modes algorithm addresses this lacuna by replacing the
Euclidean distance with the Hamming distance and the means with the modes in
the $k$-means objective function. We provide a novel, computationally efficient
implementation of $k$-modes, called OTQT. We prove that OTQT finds updates,
undetectable to existing $k$-modes algorithms, that improve the objective
function. Thus, although slightly slower per iteration owing to its algorithmic
complexity, OTQT is always more accurate per iteration and almost always faster
(and only barely slower on some datasets) to the final optimum. As a result, we
recommend OTQT as the preferred, default algorithm for all $k$-modes
implementations. We also examine five initialization methods and three types of
$K$-selection methods, many of them novel or novel applications to $k$-modes.
By examining performance on real and simulated datasets, we show that simple
random initialization is the best initializer and that a novel $K$-selection
method is more accurate than methods adapted from $k$-means.
| stat |
Use Internet Search Data to Accurately Track State-Level Influenza
Epidemics | For epidemics control and prevention, timely insights of potential hot spots
are invaluable. Alternative to traditional epidemic surveillance, which often
lags behind real time by weeks, big data from the Internet provide important
information of the current epidemic trends. Here we present a methodology,
ARGOX (Augmented Regression with GOogle data CROSS space), for accurate
real-time tracking of state-level influenza epidemics in the United States.
ARGOX combines Internet search data at the national, regional and state levels
with traditional influenza surveillance data from the Centers for Disease
Control and Prevention, and accounts for both the spatial correlation structure
of state-level influenza activities and the evolution of people's Internet
search pattern. ARGOX achieves on average 28\% error reduction over the best
alternative for real-time state-level influenza estimation for 2014 to 2020.
ARGOX is robust and reliable and can be potentially applied to track county-
and city-level influenza activity and other infectious diseases.
| stat |
Capacity Value of Solar Power and Other Variable Generation | This paper reviews methods that are used for adequacy risk assessment
considering solar power and for assessment of the capacity value of solar
power. The properties of solar power are described as seen from the perspective
of the power-system operator, comparing differences in energy availability and
capacity factors with those of wind power. Methodologies for risk calculations
considering variable generation are surveyed, including the probability
background, statistical-estimation approaches, and capacity-value metrics.
Issues in incorporating variable generation in capacity markets are described,
followed by a review of applied studies considering solar power. Finally,
recommendations for further research are presented.
| stat |
Global forensic geolocation with deep neural networks | An important problem in forensic analyses is identifying the provenance of
materials at a crime scene, such as biological material on a piece of clothing.
This procedure, known as geolocation, is conventionally guided by expert
knowledge of the biological evidence and therefore tends to be
application-specific, labor-intensive, and subjective. Purely data-driven
methods have yet to be fully realized due in part to the lack of a sufficiently
rich data source. However, high-throughput sequencing technologies are able to
identify tens of thousands of microbial taxa using DNA recovered from a single
swab collected from nearly any object or surface. We present a new algorithm
for geolocation that aggregates over an ensemble of deep neural network
classifiers trained on randomly-generated Voronoi partitions of a spatial
domain. We apply the algorithm to fungi present in each of 1300 dust samples
collected across the continental United States and then to a global dataset of
dust samples from 28 countries. Our algorithm makes remarkably good point
predictions with more than half of the geolocation errors under 100 kilometers
for the continental analysis and nearly 90% classification accuracy of a
sample's country of origin for the global analysis. We suggest that the
effectiveness of this model sets the stage for a new, quantitative approach to
forensic geolocation.
| stat |
Simultaneous Inference for Empirical Best Predictors with a Poverty
Study in Small Areas | Today, generalized linear mixed models are broadly used in many fields.
However, the development of tools for performing simultaneous inference has
been largely neglected in this domain. A framework for joint inference is
indispensable for comparisons of clusters or small areas. Therefore we
introduce simultaneous confidence intervals and develop multiple testing
procedures for empirical best predictors. We consider area- and unit-level
models to study poverty rates in the counties of Galicia in Spain. In this
context we illustrate to what extent existing methods fail. Simulation studies
and complete asymptotic theory are also provided.
| stat |
Structure-Adaptive Sequential Testing for Online False Discovery Rate
Control | Consider the online testing of a stream of hypotheses where a real--time
decision must be made before the next data point arrives. The error rate is
required to be controlled at {all} decision points. Conventional
\emph{simultaneous testing rules} are no longer applicable due to the more
stringent error constraints and absence of future data. Moreover, the online
decision--making process may come to a halt when the total error budget, or
alpha--wealth, is exhausted. This work develops a new class of
structure--adaptive sequential testing (SAST) rules for online false discover
rate (FDR) control. A key element in our proposal is a new alpha--investment
algorithm that precisely characterizes the gains and losses in sequential
decision making. SAST captures time varying structures of the data stream,
learns the optimal threshold adaptively in an ongoing manner and optimizes the
alpha-wealth allocation across different time periods. We present theory and
numerical results to show that the proposed method is valid for online FDR
control and achieves substantial power gain over existing online testing rules.
| stat |
On the Transfer of Inductive Bias from Simulation to the Real World: a
New Disentanglement Dataset | Learning meaningful and compact representations with disentangled semantic
aspects is considered to be of key importance in representation learning. Since
real-world data is notoriously costly to collect, many recent state-of-the-art
disentanglement models have heavily relied on synthetic toy data-sets. In this
paper, we propose a novel data-set which consists of over one million images of
physical 3D objects with seven factors of variation, such as object color,
shape, size and position. In order to be able to control all the factors of
variation precisely, we built an experimental platform where the objects are
being moved by a robotic arm. In addition, we provide two more datasets which
consist of simulations of the experimental setup. These datasets provide for
the first time the possibility to systematically investigate how well different
disentanglement methods perform on real data in comparison to simulation, and
how simulated data can be leveraged to build better representations of the real
world. We provide a first experimental study of these questions and our results
indicate that learned models transfer poorly, but that model and hyperparameter
selection is an effective means of transferring information to the real world.
| stat |
Forecasting unemployment using Internet search data via PRISM | Big data generated from the Internet offer great potential for predictive
analysis. Here we focus on using online users' Internet search data to forecast
unemployment initial claims weeks into the future, which provides timely
insights into the direction of the economy. To this end, we present a novel
method PRISM (Penalized Regression with Inferred Seasonality Module), which
uses publicly available online search data from Google. PRISM is a
semi-parametric method, motivated by a general state-space formulation, and
employs nonparametric seasonal decomposition and penalized regression. For
forecasting unemployment initial claims, PRISM outperforms all previously
available methods, including forecasting during the 2008-2009 financial crisis
period and near-future forecasting during the COVID-19 pandemic period, when
unemployment initial claims both rose rapidly. The timely and accurate
unemployment forecasts by PRISM could aid government agencies and financial
institutions to assess the economic trend and make well-informed decisions,
especially in the face of economic turbulence.
| stat |
Foothill: A Quasiconvex Regularization for Edge Computing of Deep Neural
Networks | Deep neural networks (DNNs) have demonstrated success for many supervised
learning tasks, ranging from voice recognition, object detection, to image
classification. However, their increasing complexity might yield poor
generalization error that make them hard to be deployed on edge devices.
Quantization is an effective approach to compress DNNs in order to meet these
constraints. Using a quasiconvex base function in order to construct a binary
quantizer helps training binary neural networks (BNNs) and adding noise to the
input data or using a concrete regularization function helps to improve
generalization error. Here we introduce foothill function, an infinitely
differentiable quasiconvex function. This regularizer is flexible enough to
deform towards $L_1$ and $L_2$ penalties. Foothill can be used as a binary
quantizer, as a regularizer, or as a loss. In particular, we show this
regularizer reduces the accuracy gap between BNNs and their full-precision
counterpart for image classification on ImageNet.
| stat |
Variance reduction for MCMC methods via martingale representations | In this paper we propose an efficient variance reduction approach for
additive functionals of Markov chains relying on a novel discrete time
martingale representation. Our approach is fully non-asymptotic and does not
require the knowledge of the stationary distribution (and even any type of
ergodicity) or specific structure of the underlying density. By rigorously
analyzing the convergence properties of the proposed algorithm, we show that
its cost-to-variance product is indeed smaller than one of the naive algorithm.
The numerical performance of the new method is illustrated for the
Langevin-type Markov Chain Monte Carlo (MCMC) methods.
| stat |
The Geometry of Nonlinear Embeddings in Kernel Discriminant Analysis | Fisher's linear discriminant analysis is a classical method for
classification, yet it is limited to capturing linear features only. Kernel
discriminant analysis as an extension is known to successfully alleviate the
limitation through a nonlinear feature mapping. We study the geometry of
nonlinear embeddings in discriminant analysis with polynomial kernels and
Gaussian kernel by identifying the population-level discriminant function that
depends on the data distribution and the kernel. In order to obtain the
discriminant function, we solve a generalized eigenvalue problem with
between-class and within-class covariance operators. The polynomial
discriminants are shown to capture the class difference through the population
moments explicitly. For approximation of the Gaussian discriminant, we use a
particular representation of the Gaussian kernel by utilizing the exponential
generating function for Hermite polynomials. We also show that the Gaussian
discriminant can be approximated using randomized projections of the data. Our
results illuminate how the data distribution and the kernel interact in
determination of the nonlinear embedding for discrimination, and provide a
guideline for choice of the kernel and its parameters.
| stat |
Random spherical hyperbolic diffusion | The paper starts by giving a motivation for this research and justifying the
considered stochastic diffusion models for cosmic microwave background
radiation studies. Then it derives the exact solution in terms of a series
expansion to a hyperbolic diffusion equation on the unit sphere. The Cauchy
problem with random initial conditions is studied. All assumptions are stated
in terms of the angular power spectrum of the initial conditions. An
approximation to the solution is given and analysed by finitely truncating the
series expansion. The upper bounds for the convergence rates of the
approximation errors are derived. Smoothness properties of the solution and its
approximation are investigated. It is demonstrated that the sample H\"older
continuity of these spherical fields is related to the decay of the angular
power spectrum. Numerical studies of approximations to the solution and
applications to cosmic microwave background data are presented to illustrate
the theoretical results.
| stat |
Distributed SGD Generalizes Well Under Asynchrony | The performance of fully synchronized distributed systems has faced a
bottleneck due to the big data trend, under which asynchronous distributed
systems are becoming a major popularity due to their powerful scalability. In
this paper, we study the generalization performance of stochastic gradient
descent (SGD) on a distributed asynchronous system. The system consists of
multiple worker machines that compute stochastic gradients which are further
sent to and aggregated on a common parameter server to update the variables,
and the communication in the system suffers from possible delays. Under the
algorithm stability framework, we prove that distributed asynchronous SGD
generalizes well given enough data samples in the training optimization. In
particular, our results suggest to reduce the learning rate as we allow more
asynchrony in the distributed system. Such adaptive learning rate strategy
improves the stability of the distributed algorithm and reduces the
corresponding generalization error. Then, we confirm our theoretical findings
via numerical experiments.
| stat |
Integrating computing in the statistics and data science curriculum:
Creative structures, novel skills and habits, and ways to teach computational
thinking | Nolan and Temple Lang (2010) argued for the fundamental role of computing in
the statistics curriculum. In the intervening decade the statistics education
community has acknowledged that computational skills are as important to
statistics and data science practice as mathematics. There remains a notable
gap, however, between our intentions and our actions. In this special issue of
the *Journal of Statistics and Data Science Education* we have assembled a
collection of papers that (1) suggest creative structures to integrate
computing, (2) describe novel data science skills and habits, and (3) propose
ways to teach computational thinking. We believe that it is critical for the
community to redouble our efforts to embrace sophisticated computing in the
statistics and data science curriculum. We hope that these papers provide
useful guidance for the community to move these efforts forward.
| stat |
Change Point Detection for Compositional Multivariate Data | Change point detection algorithms have numerous applications in fields of
scientific and economic importance. We consider the problem of change point
detection on compositional multivariate data (each sample is a probability mass
function), which is a practically important sub-class of general multivariate
data. While the problem of change-point detection is well studied in univariate
setting, and there are few viable implementations for a general multivariate
data, the existing methods do not perform well on compositional data. In this
paper, we propose a parametric approach for change point detection in
compositional data. Moreover, using simple transformations on data, we extend
our approach to handle any general multivariate data. Experimentally, we show
that our method performs significantly better on compositional data and is
competitive on general data compared to the available state of the art
implementations.
| stat |
Beyond Gaussian Approximation: Bootstrap for Maxima of Sums of
Independent Random Vectors | The Bonferroni adjustment, or the union bound, is commonly used to study rate
optimality properties of statistical methods in high-dimensional problems.
However, in practice, the Bonferroni adjustment is overly conservative. The
extreme value theory has been proven to provide more accurate multiplicity
adjustments in a number of settings, but only on ad hoc basis. Recently,
Gaussian approximation has been used to justify bootstrap adjustments in large
scale simultaneous inference in some general settings when $n \gg (\log p)^7$,
where $p$ is the multiplicity of the inference problem and $n$ is the sample
size. The thrust of this theory is the validity of the Gaussian approximation
for maxima of sums of independent random vectors in high-dimension. In this
paper, we reduce the sample size requirement to $n \gg (\log p)^5$ for the
consistency of the empirical bootstrap and the multiplier/wild bootstrap in the
Kolmogorov-Smirnov distance, possibly in the regime where the Gaussian
approximation is not available. New comparison and anti-concentration theorems,
which are of considerable interest in and of themselves, are developed as
existing ones interweaved with Gaussian approximation are no longer applicable.
| stat |
A Bayesian dose-response meta-analysis model: simulation study and
application | Dose-response models express the effect of different dose or exposure levels
on a specific outcome. In meta-analysis, where aggregated-level data is
available, dose-response evidence is synthesized using either one-stage or
two-stage models in a frequentist setting. We propose a hierarchical
dose-response model implemented in a Bayesian framework. We present the model
with cubic dose-response shapes for a dichotomous outcome and take into account
heterogeneity due to variability in the dose-response shape. We develop our
Bayesian model assuming normal or binomial likelihood and accounting for
exposures grouped in clusters. We implement these models in R using JAGS and we
compare our approach to the one-stage dose-response meta-analysis model in a
simulation study. We found that the Bayesian dose-response model with binomial
likelihood has slightly lower bias than the Bayesian model with the normal
likelihood and the frequentist one-stage model. However, all three models
perform very well and give practically identical results. We also re-analyze
the data from 60 randomized controlled trials (15,984 participants) examining
the efficacy (response) of various doses of antidepressant drugs. All models
suggest that the dose-response curve increases between zero dose and 40 mg of
fluoxetine-equivalent dose, and thereafter is constant. We draw the same
conclusion when we take into account the fact that five different
antidepressants have been studied in the included trials. We show that
implementation of the hierarchical model in Bayesian framework has similar
performance to, but overcomes some of the limitations of the frequentist
approaches and offers maximum flexibility to accommodate features of the data.
| stat |
Copula Density Estimation by Finite Mixture of Parametric Copula
Densities | A Copula density estimation method that is based on a finite mixture of
heterogeneous parametric copula densities is proposed here. More specifically,
the mixture components are Clayton, Frank, Gumbel, T, and normal copula
densities, which are capable of capturing lower tail,strong central, upper
tail, heavy tail, and symmetrical elliptical dependence, respectively. The
model parameters are estimated by an interior-point algorithm for the
constrained maximum likelihood problem. The interior-point algorithm is
compared with the commonly used EM algorithm. Simulation and real data
application show that the proposed approach is effective to model complex
dependencies for data in dimensions beyond two or three.
| stat |
Learning identifiable and interpretable latent models of
high-dimensional neural activity using pi-VAE | The ability to record activities from hundreds of neurons simultaneously in
the brain has placed an increasing demand for developing appropriate
statistical techniques to analyze such data. Recently, deep generative models
have been proposed to fit neural population responses. While these methods are
flexible and expressive, the downside is that they can be difficult to
interpret and identify. To address this problem, we propose a method that
integrates key ingredients from latent models and traditional neural encoding
models. Our method, pi-VAE, is inspired by recent progress on identifiable
variational auto-encoder, which we adapt to make appropriate for neuroscience
applications. Specifically, we propose to construct latent variable models of
neural activity while simultaneously modeling the relation between the latent
and task variables (non-neural variables, e.g. sensory, motor, and other
externally observable states). The incorporation of task variables results in
models that are not only more constrained, but also show qualitative
improvements in interpretability and identifiability. We validate pi-VAE using
synthetic data, and apply it to analyze neurophysiological datasets from rat
hippocampus and macaque motor cortex. We demonstrate that pi-VAE not only fits
the data better, but also provides unexpected novel insights into the structure
of the neural codes.
| stat |
Ordinal Probit Functional Outcome Regression with Application to
Computer-Use Behavior in Rhesus Monkeys | Research in functional regression has made great strides in expanding to
non-Gaussian functional outcomes, but exploration of ordinal functional
outcomes remains limited. Motivated by a study of computer-use behavior in
rhesus macaques (Macaca mulatta), we introduce the Ordinal Probit Functional
Outcome Regression model (OPFOR). OPFOR models can be fit using one of several
basis functions including penalized B-splines, wavelets, and O'Sullivan splines
-- the last of which typically performs best. Simulation using a variety of
underlying covariance patterns shows that the model performs reasonably well in
estimation under multiple basis functions with near nominal coverage for joint
credible intervals. Finally, in application, we use Bayesian model selection
criteria adapted to functional outcome regression to best characterize the
relation between several demographic factors of interest and the monkeys'
computer use over the course of a year. In comparison with a standard ordinal
longitudinal analysis, OPFOR outperforms a cumulative-link mixed-effects model
in simulation and provides additional and more nuanced information on the
nature of the monkeys' computer-use behavior.
| stat |
HMLasso: Lasso with High Missing Rate | Sparse regression such as the Lasso has achieved great success in handling
high-dimensional data. However, one of the biggest practical problems is that
high-dimensional data often contain large amounts of missing values. Convex
Conditioned Lasso (CoCoLasso) has been proposed for dealing with
high-dimensional data with missing values, but it performs poorly when there
are many missing values, so that the high missing rate problem has not been
resolved. In this paper, we propose a novel Lasso-type regression method for
high-dimensional data with high missing rates. We effectively incorporate mean
imputed covariance, overcoming its inherent estimation bias. The result is an
optimally weighted modification of CoCoLasso according to missing ratios. We
theoretically and experimentally show that our proposed method is highly
effective even when there are many missing values.
| stat |
Lugsail lag windows for estimating time-average covariance matrices | Lag windows are commonly used in time series, econometrics, steady-state
simulation, and Markov chain Monte Carlo to estimate time-average covariance
matrices. In the presence of high correlation, estimators of the time-average
covariance matrix almost always exhibit significant negative bias, leading to
undesirable finite-sample properties. We propose a new family of lag windows
specifically designed to improve finite-sample performance by offsetting this
negative bias in the opposite direction. Any existing lag window can be adapted
into a lugsail equivalent with no additional assumptions. We use these lag
windows within spectral variance estimators and demonstrate its advantages in a
linear regression model with autocorrelated and heteroskedastic residuals. We
further consider weighted batch means estimators since spectral variance
estimators are computationally intensive for large simulation output. We obtain
the bias and variance results for these multivariate estimators and
significantly weaken the mixing condition on the process. Superior
finite-sample properties are illustrated in a vector autoregressive process and
a Bayesian logistic regression model.
| stat |
Conflating marginal and conditional treatment effects: Comments on
'Assessing the performance of population adjustment methods for anchored
indirect comparisons: A simulation study' | In this commentary, we highlight the importance of: (1) carefully considering
and clarifying whether a marginal or conditional treatment effect is of
interest in a population-adjusted indirect treatment comparison; and (2)
developing distinct methodologies for estimating the different measures of
effect. The appropriateness of each methodology depends on the preferred target
of inference.
| stat |
PAC-Bayes under potentially heavy tails | We derive PAC-Bayesian learning guarantees for heavy-tailed losses, and
obtain a novel optimal Gibbs posterior which enjoys finite-sample excess risk
bounds at logarithmic confidence. Our core technique itself makes use of
PAC-Bayesian inequalities in order to derive a robust risk estimator, which by
design is easy to compute. In particular, only assuming that the first three
moments of the loss distribution are bounded, the learning algorithm derived
from this estimator achieves nearly sub-Gaussian statistical error, up to the
quality of the prior.
| stat |
Infinite Mixtures of Infinite Factor Analysers | Factor-analytic Gaussian mixture models are often employed as a model-based
approach to clustering high-dimensional data. Typically, the numbers of
clusters and latent factors must be specified in advance of model fitting, and
remain fixed. The pair which optimises some model selection criterion is then
chosen. For computational reasons, models in which the number of latent factors
differ across clusters are rarely considered.
Here the infinite mixture of infinite factor analysers (IMIFA) model is
introduced. IMIFA employs a Pitman-Yor process prior to facilitate automatic
inference of the number of clusters using the stick-breaking construction and a
slice sampler. Furthermore, IMIFA employs multiplicative gamma process
shrinkage priors to allow cluster-specific numbers of factors, automatically
inferred via an adaptive Gibbs sampler. IMIFA is presented as the flagship of a
family of factor-analytic mixture models, providing flexible approaches to
clustering high-dimensional data.
Applications to a benchmark data set, metabolomic spectral data, and a
manifold learning handwritten digit example illustrate the IMIFA model and its
advantageous features: IMIFA obviates the need for model selection criteria,
reduces the computational burden associated with the search of the model space,
improves clustering performance by allowing cluster-specific numbers of
factors, and quantifies uncertainty in the numbers of clusters and
cluster-specific factors.
| stat |
Integrating Markov processes with structural causal modeling enables
counterfactual inference in complex systems | This manuscript contributes a general and practical framework for casting a
Markov process model of a system at equilibrium as a structural causal model,
and carrying out counterfactual inference. Markov processes mathematically
describe the mechanisms in the system, and predict the system's equilibrium
behavior upon intervention, but do not support counterfactual inference. In
contrast, structural causal models support counterfactual inference, but do not
identify the mechanisms. This manuscript leverages the benefits of both
approaches. We define the structural causal models in terms of the parameters
and the equilibrium dynamics of the Markov process models, and counterfactual
inference flows from these settings. The proposed approach alleviates the
identifiability drawback of the structural causal models, in that the
counterfactual inference is consistent with the counterfactual trajectories
simulated from the Markov process model. We showcase the benefits of this
framework in case studies of complex biomolecular systems with nonlinear
dynamics. We illustrate that, in presence of Markov process model
misspecification, counterfactual inference leverages prior data, and therefore
estimates the outcome of an intervention more accurately than a direct
simulation.
| stat |
Gaussian Process Priors for View-Aware Inference | While frame-independent predictions with deep neural networks have become the
prominent solutions to many computer vision tasks, the potential benefits of
utilizing correlations between frames have received less attention. Even though
probabilistic machine learning provides the ability to encode correlation as
prior knowledge for inference, there is a tangible gap between the theory and
practice of applying probabilistic methods to modern vision problems. For this,
we derive a principled framework to combine information coupling between camera
poses (translation and orientation) with deep models. We proposed a novel view
kernel that generalizes the standard periodic kernel in $\mathrm{SO}(3)$. We
show how this soft-prior knowledge can aid several pose-related vision tasks
like novel view synthesis and predict arbitrary points in the latent space of
generative models, pointing towards a range of new applications for inter-frame
reasoning.
| stat |
How Robust are the Estimated Effects of Nonpharmaceutical Interventions
against COVID-19? | To what extent are effectiveness estimates of nonpharmaceutical interventions
(NPIs) against COVID-19 influenced by the assumptions our models make? To
answer this question, we investigate 2 state-of-the-art NPI effectiveness
models and propose 6 variants that make different structural assumptions. In
particular, we investigate how well NPI effectiveness estimates generalise to
unseen countries, and their sensitivity to unobserved factors. Models that
account for noise in disease transmission compare favourably. We further
evaluate how robust estimates are to different choices of epidemiological
parameters and data. Focusing on models that assume transmission noise, we find
that previously published results are remarkably robust across these variables.
Finally, we mathematically ground the interpretation of NPI effectiveness
estimates when certain common assumptions do not hold.
| stat |
On Newton Screening | Screening and working set techniques are important approaches to reducing the
size of an optimization problem. They have been widely used in accelerating
first-order methods for solving large-scale sparse learning problems. In this
paper, we develop a new screening method called Newton screening (NS) which is
a generalized Newton method with a built-in screening mechanism. We derive an
equivalent KKT system for the Lasso and utilize a generalized Newton method to
solve the KKT equations. Based on this KKT system, a built-in working set with
a relatively small size is first determined using the sum of primal and dual
variables generated from the previous iteration, then the primal variable is
updated by solving a least-squares problem on the working set and the dual
variable updated based on a closed-form expression. Moreover, we consider a
sequential version of Newton screening (SNS) with a warm-start strategy. We
show that NS possesses an optimal convergence property in the sense that it
achieves one-step local convergence. Under certain regularity conditions on the
feature matrix, we show that SNS hits a solution with the same signs as the
underlying true target and achieves a sharp estimation error bound with high
probability. Simulation studies and real data analysis support our theoretical
results and demonstrate that SNS is faster and more accurate than several
state-of-the-art methods in our comparative studies.
| stat |
Missed opportunities in large scale comparison of QSAR and conformal
prediction methods and their applications in drug discovery | Recently Bosc et al. (J Cheminform 11(1): 4, 2019), published an article
describing a case study that directly compares conformal predictions with
traditional QSAR methods for large-scale predictions of target-ligand binding.
We consider this study to be very important. Unfortunately, we have found
several issues in the authors' approach as well as in the presentation of their
findings.
| stat |
Unified Robust Boosting | Boosting is a popular machine learning algorithm in regression and
classification problems. Boosting can combine a sequence of regression trees to
obtain accurate prediction. In the presence of outliers, traditional boosting,
based on optimizing convex loss functions, may show inferior results. In this
article, a unified robust boosting is proposed for more resistant estimation.
The method utilizes a recently developed concave-convex family for robust
estimation, composite optimization by conjugation operator, and functional
decent boosting. As a result, an iteratively reweighted boosting algorithm can
be conveniently constructed with existing software. Applications in robust
regression, classification and Poisson regression are demonstrated in the R
package ccboost.
| stat |
Estimating and Projecting Air Passenger Traffic during the COVID-19
Coronavirus Outbreak and its Socio-Economic Impact | The main focus of this study is to collect and prepare data on air passengers
traffic worldwide with the scope of analyze the impact of travel ban on the
aviation sector. Based on historical data from January 2010 till October 2019,
a forecasting model is implemented in order to set a reference baseline. Making
use of airplane movements extracted from online flight tracking platforms and
on-line booking systems, this study presents also a first assessment of recent
changes in flight activity around the world as a result of the COVID-19
pandemic. To study the effects of air travel ban on aviation and in turn its
socio-economic, several scenarios are constructed based on past pandemic crisis
and the observed flight volumes. It turns out that, according to this
hypothetical scenarios, in the first Quarter of 2020 the impact of aviation
losses could have negatively reduced World GDP by 0.02% to 0.12% according to
the observed data and, in the worst case scenarios, at the end of 2020 the loss
could be as high as 1.41-1.67% and job losses may reach the value of 25-30
millions. Focusing on EU27, the GDP loss may amount to 1.66-1.98% by the end of
2020 and the number of job losses from 4.2 to 5 millions in the worst case
scenarios. Some countries will be more affected than others in the short run
and most European airlines companies will suffer from the travel ban.
| stat |
Consistency between household and county measures of K-12 onsite
schooling during the COVID-19 pandemic | The academic, socioemotional, and health impacts of school policies
throughout the COVID-19 pandemic have been a source of many important questions
that require accurate information about the extent of onsite schooling that has
been occurring throughout the pandemic. This paper investigates school
operational status data sources during the COVID-19 pandemic, comparing
self-report data collected nationally on the household level through a
Facebook-based survey with data collected at district and county levels
throughout the country. The percentage of households reporting in-person
instruction within each county is compared to the district and county data at
the state and county levels. The results show high levels of consistency
between the sources at the state level and for large counties. The consistency
levels across sources support the usage of the Facebook-based COVID-19 Symptom
Survey as a source to answer questions about the educational experiences,
factors, and impacts related to K-12 education across the nation during the
pandemic.
| stat |
Thirty Years of The Network Scale up Method | Estimating the size of hard-to-reach populations is an important problem for
many fields. The Network Scale-up Method (NSUM) is a relatively new approach to
estimate the size of these hard-to-reach populations by asking respondents the
question, "How many X's do you know," where X is the population of interest
(e.g. "How many female sex workers do you know?"). The answers to these
questions form Aggregated Relational Data (ARD). The NSUM has been used to
estimate the size of a variety of subpopulations, including female sex workers,
drug users, and even children who have been hospitalized for choking. Within
the Network Scale-up methodology, there are a multitude of estimators for the
size of the hidden population, including direct estimators, maximum likelihood
estimators, and Bayesian estimators. In this article, we first provide an
in-depth analysis of ARD properties and the techniques to collect the data.
Then, we comprehensively review different estimation methods in terms of the
assumptions behind each model, the relationships between the estimators, and
the practical considerations of implementing the methods. Finally, we provide a
summary of the dominant methods and an extensive list of the applications, and
discuss the open problems and potential research directions in this area.
| stat |
A Classical Search Game in Discrete Locations | Consider a two-person zero-sum search game between a hider and a searcher.
The hider hides among $n$ discrete locations, and the searcher successively
visits individual locations until finding the hider. Known to both players, a
search at location $i$ takes $t_i$ time units and detects the hider -- if
hidden there -- independently with probability $q_i$, for $i=1,\ldots,n$. The
hider aims to maximize the expected time until detection, while the searcher
aims to minimize it. We prove the existence of an optimal strategy for each
player. In particular, the hider's optimal mixed strategy hides in each
location with a nonzero probability, and the searcher's optimal mixed strategy
can be constructed with up to $n$ simple search sequences. We develop an
algorithm to compute an optimal strategy for each player, and compare the
optimal hiding strategy with the simple hiding strategy which gives the
searcher no location preference at the beginning of the search.
| stat |
Spatial Modeling for Correlated Cancers Using Bivariate Directed Graphs | Disease maps are an important tool in cancer epidemiology used for the
analysis of geographical variations in disease rates and the investigation of
environmental risk factors underlying spatial patterns. Cancer maps help
epidemiologists highlight geographic areas with high and low prevalence,
incidence, or mortality rates of cancers, and the variability of such rates
over a spatial domain. When more than one cancer is of interest, the models
must also capture the inherent or endemic association between the diseases in
addition to the spatial association. This article develops interpretable and
easily implementable spatial autocorrelation models for two or more cancers.
The article builds upon recent developments in univariate disease mapping that
have shown the use of mathematical structures such as directed acyclic graphs
to capture spatial association for a single cancer, estimating inherent or
endemic association for two cancers in addition to the association over space
(clustering) for each of the cancers. The method builds a Bayesian hierarchical
model where the spatial effects are introduced as latent random effects for
each cancer. We analyze the relationship between incidence rates of esophagus
and lung cancer extracted from the Surveillance, Epidemiology, and End Results
(SEER) Program. Our analysis shows statistically significant association
between the county-wise incidence rates of lung and esophagus cancer across
California. The bivariate directed acyclic graphical model performs better than
competing bivariate spatial models in the existing literature.
| stat |
An Alternative Probabilistic Interpretation of the Huber Loss | The Huber loss is a robust loss function used for a wide range of regression
tasks. To utilize the Huber loss, a parameter that controls the transitions
from a quadratic function to an absolute value function needs to be selected.
We believe the standard probabilistic interpretation that relates the Huber
loss to the Huber density fails to provide adequate intuition for identifying
the transition point. As a result, a hyper-parameter search is often necessary
to determine an appropriate value. In this work, we propose an alternative
probabilistic interpretation of the Huber loss, which relates minimizing the
loss to minimizing an upper-bound on the Kullback-Leibler divergence between
Laplace distributions, where one distribution represents the noise in the
ground-truth and the other represents the noise in the prediction. In addition,
we show that the parameters of the Laplace distributions are directly related
to the transition point of the Huber loss. We demonstrate, through a toy
problem, that the optimal transition point of the Huber loss is closely related
to the distribution of the noise in the ground-truth data. As a result, our
interpretation provides an intuitive way to identify well-suited
hyper-parameters by approximating the amount of noise in the data, which we
demonstrate through a case study and experimentation on the Faster R-CNN and
RetinaNet object detectors.
| stat |
Combating Label Noise in Deep Learning Using Abstention | We introduce a novel method to combat label noise when training deep neural
networks for classification. We propose a loss function that permits abstention
during training thereby allowing the DNN to abstain on confusing samples while
continuing to learn and improve classification performance on the non-abstained
samples. We show how such a deep abstaining classifier (DAC) can be used for
robust learning in the presence of different types of label noise. In the case
of structured or systematic label noise -- where noisy training labels or
confusing examples are correlated with underlying features of the data--
training with abstention enables representation learning for features that are
associated with unreliable labels. In the case of unstructured (arbitrary)
label noise, abstention during training enables the DAC to be used as an
effective data cleaner by identifying samples that are likely to have label
noise. We provide analytical results on the loss function behavior that enable
dynamic adaption of abstention rates based on learning progress during
training. We demonstrate the utility of the deep abstaining classifier for
various image classification tasks under different types of label noise; in the
case of arbitrary label noise, we show significant improvements over previously
published results on multiple image benchmarks. Source code is available at
https://github.com/thulas/dac-label-noise
| stat |
Parallel integrative learning for large-scale multi-response regression
with incomplete outcomes | Multi-task learning is increasingly used to investigate the association
structure between multiple responses and a single set of predictor variables in
many applications. In the era of big data, the coexistence of incomplete
outcomes, large number of responses, and high dimensionality in predictors
poses unprecedented challenges in estimation, prediction, and computation. In
this paper, we propose a scalable and computationally efficient procedure,
called PEER, for large-scale multi-response regression with incomplete
outcomes, where both the numbers of responses and predictors can be
high-dimensional. Motivated by sparse factor regression, we convert the
multi-response regression into a set of univariate-response regressions, which
can be efficiently implemented in parallel. Under some mild regularity
conditions, we show that PEER enjoys nice sampling properties including
consistency in estimation, prediction, and variable selection. Extensive
simulation studies show that our proposal compares favorably with several
existing methods in estimation accuracy, variable selection, and computation
efficiency.
| stat |
Statistical Models in Forensic Voice Comparison | This chapter describes a number of signal-processing and statistical-modeling
techniques that are commonly used to calculate likelihood ratios in
human-supervised automatic approaches to forensic voice comparison. Techniques
described include mel-frequency cepstral coefficients (MFCCs) feature
extraction, Gaussian mixture model - universal background model (GMM-UBM)
systems, i-vector - probabilistic linear discriminant analysis (i-vector PLDA)
systems, deep neural network (DNN) based systems (including senone posterior
i-vectors, bottleneck features, and embeddings / x-vectors), mismatch
compensation, and score-to-likelihood-ratio conversion (aka calibration).
Empirical validation of forensic-voice-comparison systems is also covered. The
aim of the chapter is to bridge the gap between general introductions to
forensic voice comparison and the highly technical
automatic-speaker-recognition literature from which the signal-processing and
statistical-modeling techniques are mostly drawn. Knowledge of the
likelihood-ratio framework for the evaluation of forensic evidence is assumed.
It is hoped that the material presented here will be of value to students of
forensic voice comparison and to researchers interested in learning about
statistical modeling techniques that could potentially also be applied to data
from other branches of forensic science.
| stat |
Analysis of the fiber laydown quality in spunbond processes with
simulation experiments evaluated by blocked neural networks | We present a simulation framework for spunbond processes and use a design of
experiments to investigate the cause-and-effect-relations of process and
material parameters onto the fiber laydown on a conveyor belt. The virtual
experiments are analyzed by a blocked neural network. This forms the basis for
the prediction of the fiber laydown characteristics and enables a quick ranking
of the significance of the influencing effects. We conclude our research by an
analysis of the nonlinear cause-and-effect relations.
| stat |
Naive linkage error corrected dual system estimation | The utility of capture-recapture methods is offset by strong underlying
assumptions. In this paper we discuss a failure in the perfect linkage
assumption in a special case of capture-recapture known as the dual system
estimation. We propose a naive linkage error corrected dual system estimator.
The advantage of the proposed approach compared with the existing linkage error
corrected methods is that it permits an easy way to obtain the corresponding
variance estimator.
| stat |
A framework for causal inference in the presence of extreme inverse
probability weights: the role of overlap weights | In this paper, we consider recent progress in estimating the average
treatment effect when extreme inverse probability weights are present and focus
on methods that account for a possible violation of the positivity assumption.
These methods aim at estimating the treatment effect on the subpopulation of
patients for whom there is a clinical equipoise. We propose a systematic
approach to determine their related causal estimands and develop new insights
into the properties of the weights targeting such a subpopulation. Then, we
examine the roles of overlap weights, matching weights, Shannon's entropy
weights, and beta weights. This helps us characterize and compare their
underlying estimators, analytically and via simulations, in terms of the
accuracy, precision, and root mean squared error. Moreover, we study the
asymptotic behaviors of their augmented estimators (that mimic doubly robust
estimators), which lead to improved estimations when either the propensity or
the regression models are correctly specified. Based on the analytical and
simulation results, we conclude that overall overlap weights are preferable to
matching weights, especially when there is moderate or extreme violations of
the positivity assumption. Finally, we illustrate the methods using a real data
example marked by extreme inverse probability weights.
| stat |
Optimized Auxiliary Particle Filters | Auxiliary particle filters (APFs) are a class of sequential Monte Carlo (SMC)
methods for Bayesian inference in state-space models. In their original
derivation, APFs operate in an extended space using an auxiliary variable to
improve the inference. Later works have re-interpreted APFs from a multiple
importance sampling perspective. In this perspective, the proposal is a mixture
composed of kernels and weights that are selected by taking into account the
latest observation. In this work, we further exploit this perspective by
proposing an online, flexible framework for APFs that adapts the mixture
proposal by convex optimization and allows for a controllable computational
complexity. We minimize the discrepancy between the proposal and the filtering
distribution at a set of relevant points, which are chosen by leveraging the
structure of SMC. We compare our method to state-of-the-art particle filters,
showing better performance in challenging and widely used dynamical models.
| stat |
Boosting in Univariate Nonparametric Maximum Likelihood Estimation | Nonparametric maximum likelihood estimation is intended to infer the unknown
density distribution while making as few assumptions as possible. To alleviate
the over parameterization in nonparametric data fitting, smoothing assumptions
are usually merged into the estimation. In this paper a novel boosting-based
method is introduced to the nonparametric estimation in univariate cases. We
deduce the boosting algorithm by the second-order approximation of
nonparametric log-likelihood. Gaussian kernel and smooth spline are chosen as
weak learners in boosting to satisfy the smoothing assumptions. Simulations and
real data experiments demonstrate the efficacy of the proposed approach.
| stat |
Random deep neural networks are biased towards simple functions | We prove that the binary classifiers of bit strings generated by random wide
deep neural networks with ReLU activation function are biased towards simple
functions. The simplicity is captured by the following two properties. For any
given input bit string, the average Hamming distance of the closest input bit
string with a different classification is at least sqrt(n / (2{\pi} log n)),
where n is the length of the string. Moreover, if the bits of the initial
string are flipped randomly, the average number of flips required to change the
classification grows linearly with n. These results are confirmed by numerical
experiments on deep neural networks with two hidden layers, and settle the
conjecture stating that random deep neural networks are biased towards simple
functions. This conjecture was proposed and numerically explored in [Valle
P\'erez et al., ICLR 2019] to explain the unreasonably good generalization
properties of deep learning algorithms. The probability distribution of the
functions generated by random deep neural networks is a good choice for the
prior probability distribution in the PAC-Bayesian generalization bounds. Our
results constitute a fundamental step forward in the characterization of this
distribution, therefore contributing to the understanding of the generalization
properties of deep learning algorithms.
| stat |
D-STEM v2: A Software for Modelling Functional Spatio-Temporal Data | Functional spatio-temporal data naturally arise in many environmental and
climate applications where data are collected in a three-dimensional space over
time. The MATLAB D-STEM v1 software package was first introduced for modelling
multivariate space-time data and has been recently extended to D-STEM v2 to
handle functional data indexed across space and over time. This paper
introduces the new modelling capabilities of D-STEM v2 as well as the
complexity reduction techniques required when dealing with large data sets.
Model estimation, validation and dynamic kriging are demonstrated in two case
studies, one related to ground-level air quality data in Beijing, China, and
the other one related to atmospheric profile data collected globally through
radio sounding.
| stat |
Curb Your Normality: On the Quality Requirements of Demand Prediction
for Dynamic Public Transport | As Public Transport (PT) becomes more dynamic and demand-responsive, it
increasingly depends on predictions of transport demand. But how accurate need
such predictions be for effective PT operation? We address this question
through an experimental case study of PT trips in Metropolitan Copenhagen,
Denmark, which we conduct independently of any specific prediction models.
First, we simulate errors in demand prediction through unbiased noise
distributions that vary considerably in shape. Using the noisy predictions, we
then simulate and optimize demand-responsive PT fleets via a commonly used
linear programming formulation and measure their performance. Our results
suggest that the optimized performance is mainly affected by the skew of the
noise distribution and the presence of infrequently large prediction errors. In
particular, the optimized performance can improve under non-Gaussian vs.
Gaussian noise. We also obtain that dynamic routing can reduce trip time by at
least 23% vs. static routing. This reduction is estimated at 809,000 EUR per
year in terms of Value of Travel Time Savings for the case study.
| stat |
Estimating heterogeneous effects of continuous exposures using Bayesian
tree ensembles: revisiting the impact of abortion rates on crime | In estimating the causal effect of a continuous exposure or treatment, it is
important to control for all confounding factors. However, most existing
methods require parametric specification for how control variables influence
the outcome or generalized propensity score, and inference on treatment effects
is usually sensitive to this choice. Additionally, it is often the goal to
estimate how the treatment effect varies across observed units. To address this
gap, we propose a semiparametric model using Bayesian tree ensembles for
estimating the causal effect of a continuous treatment of exposure which (i)
does not require a priori parametric specification of the influence of control
variables, and (ii) allows for identification of effect modification by
pre-specified moderators. The main parametric assumption we make is that the
effect of the exposure on the outcome is linear, with the steepness of this
relationship determined by a nonparametric function of the moderators, and we
provide heuristics to diagnose the validity of this assumption. We apply our
methods to revisit a 2001 study of how abortion rates affect incidence of
crime.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.