title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Space Partitioning and Regression Mode Seeking via a Mean-Shift-Inspired
Algorithm | The mean shift (MS) algorithm is a nonparametric method used to cluster
sample points and find the local modes of kernel density estimates, using an
idea based on iterative gradient ascent. In this paper we develop a
mean-shift-inspired algorithm to estimate the modes of regression functions and
partition the sample points in the input space. We prove convergence of the
sequences generated by the algorithm and derive the non-asymptotic rates of
convergence of the estimated local modes for the underlying regression model.
We also demonstrate the utility of the algorithm for data-enabled discovery
through an application on biomolecular structure data. An extension to subspace
constrained mean shift (SCMS) algorithm used to extract ridges of regression
functions is briefly discussed.
| stat |
Geographic ratemaking with spatial embeddings | Spatial data is a rich source of information for actuarial applications:
knowledge of a risk's location could improve an insurance company's ratemaking,
reserving or risk management processes. Insurance companies with high exposures
in a territory typically have a competitive advantage since they may use
historical losses in a region to model spatial risk non-parametrically. Relying
on geographic losses is problematic for areas where past loss data is
unavailable. This paper presents a method based on data (instead of smoothing
historical insurance claim losses) to construct a geographic ratemaking model.
In particular, we construct spatial features within a complex representation
model, then use the features as inputs to a simpler predictive model (like a
generalized linear model). Our approach generates predictions with smaller bias
and smaller variance than other spatial interpolation models such as bivariate
splines in most situations. This method also enables us to generate rates in
territories with no historical experience.
| stat |
Functional time series forecasting of extreme values | We consider forecasting functional time series of extreme values within a
generalised extreme value distribution (GEV). The GEV distribution can be
characterised using the three parameters (location, scale and shape). As a
result, the forecasts of the GEV density can be accomplished by forecasting
these three latent parameters. Depending on the underlying data structure, some
of the three parameters can either be modelled as scalars or functions. We
provide two forecasting algorithms to model and forecast these parameters. To
assess the forecast uncertainty, we apply a sieve bootstrap method to construct
pointwise and simultaneous prediction intervals of the forecasted extreme
values. Illustrated by a daily maximum temperature dataset, we demonstrate the
advantages of modelling these parameters as functions. Further, the
finite-sample performance of our methods is quantified using several
Monte-Carlo simulated data under a range of scenarios.
| stat |
Notes on Exact Power Calculations for t Tests and Analysis of Covariance | Tang derived the exact power formulae for t tests and analysis of covariance
(ANCOVA) in superiority, noninferiority and equivalence trials. The power
calculation in equivalence trials can be simplified by using Owen's Q function,
which is available in standard statistical software. We extend the exact power
determination method for ANCOVA to unstratified and stratified multi-arm
randomized trials. The method is applied to the design of multi-arm trials and
gold standard noninferiority trials.
| stat |
Adaptive Frequency Band Analysis for Functional Time Series | The frequency-domain properties of nonstationary functional time series often
contain valuable information. These properties are characterized through its
time-varying power spectrum. Practitioners seeking low-dimensional summary
measures of the power spectrum often partition frequencies into bands and
create collapsed measures of power within bands. However, standard frequency
bands have largely been developed through manual inspection of time series data
and may not adequately summarize power spectra. In this article, we propose a
framework for adaptive frequency band estimation of nonstationary functional
time series that optimally summarizes the time-varying dynamics of the series.
We develop a scan statistic and search algorithm to detect changes in the
frequency domain. We establish theoretical properties of this framework and
develop a computationally-efficient implementation. The validity of our method
is also justified through numerous simulation studies and an application to
analyzing electroencephalogram data in participants alternating between eyes
open and eyes closed conditions.
| stat |
Identifying relationships between cognitive processes across tasks,
contexts, and time | It is commonly assumed that a specific testing occasion (task, design,
procedure, etc.) provides insights that generalise beyond that occasion. This
assumption is infrequently carefully tested in data. We develop a statistically
principled method to directly estimate the correlation between latent
components of cognitive processing across tasks, contexts, and time. This
method simultaneously estimates individual-participant parameters of a
cognitive model at each testing occasion, group-level parameters representing
across-participant parameter averages and variances, and across-task
correlations. The approach provides a natural way to "borrow" strength across
testing occasions, which can increase the precision of parameter estimates
across all testing occasions. Two example applications demonstrate that the
method is practical in standard designs. The examples, and a simulation study,
also provide evidence about the reliability and validity of parameter estimates
from the linear ballistic accumulator model. We conclude by highlighting the
potential of the parameter-correlation method to provide an "assumption-light"
tool for estimating the relatedness of cognitive processes across tasks,
contexts, and time.
| stat |
The Mathematics Behind Spectral Clustering And The Equivalence To PCA | Spectral clustering is a popular algorithm that clusters points using the
eigenvalues and eigenvectors of Laplacian matrices derived from the data. For
years, spectral clustering has been working mysteriously. This paper explains
spectral clustering by dividing it into two categories based on whether the
graph Laplacian is fully connected or not. For a fully connected graph, this
paper demonstrates the dimension reduction part by offering an objective
function: the covariance between the original data points' similarities and the
mapped data points' similarities. For a multi-connected graph, this paper
proves that with a proper $k$, the first $k$ eigenvectors are the indicators of
the connected components. This paper also proves there is an equivalence
between spectral embedding and PCA.
| stat |
A modified weighted log-rank test for confirmatory trials with a high
proportion of treatment switching | In confirmatory cancer clinical trials, overall survival (OS) is normally a
primary endpoint in the intention-to-treat (ITT) analysis under regulatory
standards. After the tumor progresses, it is common that patients allocated to
the control group switch to the experimental treatment, or another drug in the
same class. Such treatment switching may dilute the relative efficacy of the
new drug compared to the control group, leading to lower statistical power. It
would be possible to decrease the estimation bias by shortening the follow-up
period but this may lead to a loss of information and power. Instead we propose
a modified weighted log-rank test (mWLR) that aims at balancing these factors
by down-weighting events occurring when many patients have switched treatment.
As the weighting should be pre-specified and the impact of treatment
switching is unknown, we predict the hazard ratio function and use it to
compute the weights of the mWLR. The method may incorporate information from
previous trials regarding the potential hazard ratio function over time.
We are motivated by the RECORD-1 trial of everolimus against placebo in
patients with metastatic renal-cell carcinoma where almost 80\% of the patients
in the placebo group received everolimus after disease progression. Extensive
simulations show that the new test gives considerably higher efficiency than
the standard log-rank test in realistic scenarios.
| stat |
$M$-estimation in a diffusion model with application to biosensor
transdermal blood alcohol monitoring | With the goal of well-founded statistical inference on an individual's blood
alcohol level based on noisy measurements of their skin alcohol content, we
develop $M$-estimation methodology in a general setting. We then apply it to a
diffusion equation-based model for the blood/skin alcohol relationship thereby
establishing existence, consistency, and asymptotic normality of the nonlinear
least squares estimator of the diffusion model's parameter. Simulation studies
show agreement between the estimator's performance and its asymptotic
distribution, and it is applied to a real skin alcohol data set collected via
biosensor.
| stat |
Privacy-Preserving Distributed SVD via Federated Power | Singular value decomposition (SVD) is one of the most fundamental tools in
machine learning and statistics.The modern machine learning community usually
assumes that data come from and belong to small-scale device users. The low
communication and computation power of such devices, and the possible privacy
breaches of users' sensitive data make the computation of SVD challenging.
Federated learning (FL) is a paradigm enabling a large number of devices to
jointly learn a model in a communication-efficient way without data sharing. In
the FL framework, we develop a class of algorithms called FedPower for the
computation of partial SVD in the modern setting. Based on the well-known power
method, the local devices alternate between multiple local power iterations and
one global aggregation to improve communication efficiency. In the aggregation,
we propose to weight each local eigenvector matrix with Orthogonal Procrustes
Transformation (OPT). Considering the practical stragglers' effect, the
aggregation can be fully participated or partially participated, where for the
latter we propose two sampling and aggregation schemes. Further, to ensure
strong privacy protection, we add Gaussian noise whenever the communication
happens by adopting the notion of differential privacy (DP). We theoretically
show the convergence bound for FedPower. The resulting bound is interpretable
with each part corresponding to the effect of Gaussian noise, parallelization,
and random sampling of devices, respectively. We also conduct experiments to
demonstrate the merits of FedPower. In particular, the local iterations not
only improve communication efficiency but also reduce the chance of privacy
breaches.
| stat |
Joint Mean-Covariance Estimation via the Horseshoe with an Application
in Genomic Data Analysis | Seemingly unrelated regression is a natural framework for regressing multiple
correlated responses on multiple predictors. The model is very flexible, with
multiple linear regression and covariance selection models being special cases.
However, its practical deployment in genomic data analysis under a Bayesian
framework is limited due to both statistical and computational challenges. The
statistical challenge is that one needs to infer both the mean vector and the
inverse covariance matrix, a problem inherently more complex than separately
estimating each. The computational challenge is due to the dimensionality of
the parameter space that routinely exceeds the sample size. We propose the use
of horseshoe priors on both the mean vector and the inverse covariance matrix.
This prior has demonstrated excellent performance when estimating a mean vector
or inverse covariance matrix separately. The current work shows these
advantages are also present when addressing both simultaneously. A full
Bayesian treatment is proposed, with a sampling algorithm that is linear in the
number of predictors. MATLAB code implementing the algorithm is freely
available from github at https://github.com/liyf1988/HS_GHS. Extensive
performance comparisons are provided with both frequentist and Bayesian
alternatives, and both estimation and prediction performances are verified on a
genomic data set.
| stat |
Bayes Optimal Informer Sets for Early-Stage Drug Discovery | An important experimental design problem in early-stage drug discovery is how
to prioritize available compounds for testing when very little is known about
the target protein. Informer based ranking (IBR) methods address the
prioritization problem when the compounds have provided bioactivity data on
other potentially relevant targets. An IBR method selects an informer set of
compounds, and then prioritizes the remaining compounds on the basis of new
bioactivity experiments performed with the informer set on the target. We
formalize the problem as a two-stage decision problem and introduce the Bayes
Optimal Informer SEt (BOISE) method for its solution. BOISE leverages a
flexible model of the initial bioactivity data, a relevant loss function, and
effective computational schemes to resolve the two-step design problem. We
evaluate BOISE and compare it to other IBR strategies in two retrospective
studies, one on protein-kinase inhibition and the other on anti-cancer drug
sensitivity. In both empirical settings BOISE exhibits better predictive
performance than available methods. It also behaves well with missing data,
where methods that use matrix completion show worse predictive performance. We
provide an R implementation of BOISE at
https://github.com/wiscstatman/esdd/BOISE
| stat |
Sparse Nonnegative Tensor Factorization and Completion with Noisy
Observations | In this paper, we study the sparse nonnegative tensor factorization and
completion problem from partial and noisy observations for third-order tensors.
Because of sparsity and nonnegativity, the underling tensor is decomposed into
the tensor-tensor product of one sparse nonnegative tensor and one nonnegative
tensor. We propose to minimize the sum of the maximum likelihood estimate for
the observations with nonnegativity constraints and the tensor $\ell_0$ norm
for the sparse factor. We show that the error bounds of the estimator of the
proposed model can be established under general noise observations. The
detailed error bounds under specific noise distributions including additive
Gaussian noise, additive Laplace noise, and Poisson observations can be
derived. Moreover, the minimax lower bounds are shown to be matched with the
established upper bounds up to a logarithmic factor of the sizes of the
underlying tensor. These theoretical results for tensors are better than those
obtained for matrices, and this illustrates the advantage of the use of
nonnegative sparse tensor models for completion and denoising. Numerical
experiments are provided to validate the superiority of the proposed
tensor-based method compared with the matrix-based approach.
| stat |
An Intrinsic Geometrical Approach for Statistical Process Control of
Surface and Manifold Data | We present a new method for statistical process control (SPC) of a discrete
part manufacturing system based on intrinsic geometrical properties of the
parts, estimated from three-dimensional sensor data. An intrinsic method has
the computational advantage of avoiding the difficult part registration
problem, necessary in previous SPC approaches of three-dimensional geometrical
data, but inadequate if noncontact sensors are used. The approach estimates the
spectrum of the Laplace-Beltrami (LB) operator of the scanned parts and uses a
multivariate nonparametric control chart for online process control. Our
proposal brings SPC closer to computer vision and computer graphics methods
aimed to detect large differences in shape (but not in size). However, the SPC
problem differs in that small changes in either shape or size of the parts need
to be detected, keeping a controllable false alarm rate and without completely
filtering noise. An online or "Phase II" method and a scheme for starting up in
the absence of prior data ("Phase I") are presented. Comparison with earlier
approaches that require registration shows the LB spectrum method to be more
sensitive to rapidly detect small changes in shape and size, including the
practical case when the sequence of part datasets is in the form of large,
unequal size meshes. A post-alarm diagnostic method to investigate the location
of defects on the surface of a part is also presented. While we focus in this
article on surface (triangulation) data, the methods can also be applied to
point cloud and voxel metrology data.
| stat |
Adjusting the Benjamini-Hochberg method for controlling the false
discovery rate in knockoff assisted variable selection | This paper revisits the knockoff-based multiple testing setup considered in
Barber & Candes (2015) for variable selection applied to a linear regression
model with $n\ge 2d$, where $n$ is the sample size and $d$ is the number of
explanatory variables. The BH method based on ordinary least squares estimates
of the regressions coefficients is adjusted to this setup, making it a valid
$p$-value based FDR controlling method that does not rely on any specific
correlation structure of the explanatory variables. Simulations and real data
applications demonstrate that our proposed method in its original form and its
data-adaptive version incorporating estimated proportion of truly unimportant
explanatory variables are powerful competitors of the FDR controlling methods
in Barber & Candes (2015).
| stat |
An extension of the angular synchronization problem to the heterogeneous
setting | Given an undirected measurement graph $G = ([n], E)$, the classical angular
synchronization problem consists of recovering unknown angles
$\theta_1,\dots,\theta_n$ from a collection of noisy pairwise measurements of
the form $(\theta_i - \theta_j) \mod 2\pi$, for each $\{i,j\} \in E$. This
problem arises in a variety of applications, including computer vision, time
synchronization of distributed networks, and ranking from preference
relationships. In this paper, we consider a generalization to the setting where
there exist $k$ unknown groups of angles $\theta_{l,1}, \dots,\theta_{l,n}$,
for $l=1,\dots,k$. For each $ \{i,j\} \in E$, we are given noisy pairwise
measurements of the form $\theta_{\ell,i} - \theta_{\ell,j}$ for an unknown
$\ell \in \{1,2,\ldots,k\}$. This can be thought of as a natural extension of
the angular synchronization problem to the heterogeneous setting of multiple
groups of angles, where the measurement graph has an unknown edge-disjoint
decomposition $G = G_1 \cup G_2 \ldots \cup G_k$, where the $G_i$'s denote the
subgraphs of edges corresponding to each group. We propose a probabilistic
generative model for this problem, along with a spectral algorithm for which we
provide a detailed theoretical analysis in terms of robustness against both
sampling sparsity and noise. The theoretical findings are complemented by a
comprehensive set of numerical experiments, showcasing the efficacy of our
algorithm under various parameter regimes. Finally, we consider an application
of bi-synchronization to the graph realization problem, and provide along the
way an iterative graph disentangling procedure that uncovers the subgraphs
$G_i$, $i=1,\ldots,k$ which is of independent interest, as it is shown to
improve the final recovery accuracy across all the experiments considered.
| stat |
False Discovery Rate Computation: Illustrations and Modifications | False discovery rates (FDR) are an essential component of statistical
inference, representing the propensity for an observed result to be mistaken.
FDR estimates should accompany observed results to help the user contextualize
the relevance and potential impact of findings. This paper introduces a new
user-friendly R package for computing FDRs and adjusting p-values for FDR
control. These tools respect the critical difference between the adjusted
p-value and the estimated FDR for a particular finding, which are sometimes
numerically identical but are often confused in practice. Newly augmented
methods for estimating the null proportion of findings - an important part of
the FDR estimation procedure - are proposed and evaluated. The package is
broad, encompassing a variety of methods for FDR estimation and FDR control,
and includes plotting functions for easy display of results. Through extensive
illustrations, we strongly encourage wider reporting of false discovery rates
for observed findings.
| stat |
Center-specific causal inference with multicenter trials: reinterpreting
trial evidence in the context of each participating center | In multicenter randomized trials, when effect modifiers have a different
distribution across centers, comparisons between treatment groups that average
over centers may not apply to any of the populations underlying the individual
centers. Here, we describe methods for reinterpreting the evidence produced by
a multicenter trial in the context of the population underlying each center. We
describe how to identify center-specific effects under identifiability
conditions that are largely supported by the study design and when associations
between center membership and the outcome may be present, given baseline
covariates and treatment ("center-outcome associations"). We then consider an
additional condition of no center-outcome associations given baseline
covariates and treatment. We show that this condition can be assessed using the
trial data; when it holds, center-specific treatment effects can be estimated
using analyses that completely pool information across centers. We propose
methods for estimating center-specific average treatment effects, when
center-outcome associations may be present and when they are absent, and
describe approaches for assessing whether center-specific treatment effects are
homogeneous. We evaluate the performance of the methods in a simulation study
and illustrate their implementation using data from the Hepatitis C Antiviral
Long-Term Treatment Against Cirrhosis trial.
| stat |
Bayesian Nonparametric Multivariate Spatial Mixture Mixed Effects Models
with Application to American Community Survey Special Tabulations | Leveraging multivariate spatial dependence to improve the precision of
estimates using American Community Survey data and other sample survey data has
been a topic of recent interest among data-users and federal statistical
agencies. One strategy is to use a multivariate spatial mixed effects model
with a Gaussian observation model and latent Gaussian process model. In
practice, this works well for a wide range of tabulations. Nevertheless, in
situations that exhibit heterogeneity among geographies and/or sparsity in the
data, the Gaussian assumptions may be problematic and lead to underperformance.
To remedy these situations, we propose a multivariate hierarchical Bayesian
nonparametric mixed effects spatial mixture model to increase model
flexibility. The number of clusters is chosen automatically in a data-driven
manner. The effectiveness of our approach is demonstrated through a simulation
study and motivating application of special tabulations for American Community
Survey data.
| stat |
Using Learning Dynamics to Explore the Role of Implicit Regularization
in Adversarial Examples | Recent work (Ilyas et al., 2019) suggests that adversarial examples are
features not bugs. If adversarial perturbations are indeed useful but
non-robust features, what is their origin? To answer this question, we
performed a novel analysis of the learning dynamics of adversarial
perturbations, both in pixel and frequency domains, and a systematic
steganography experiment to explore the implicit bias induced by different
model parametrizations. We find that: (1) adversarial examples are not present
at initialization but instead emerge during training; (2) the frequency-based
nature of common adversarial perturbations in natural images is critically
dependent on an implicit bias towards L1-sparsity in the frequency domain; and
(3) the origin of this bias is the locality and translation invariance of
convolutional filters, along with (4) the existence of useful frequency-based
features in the datasets. We propose a simple theoretical explanation for these
findings, providing a clear and minimalist target for theorists in future work.
Looking forward, our work shows that analyzing the learning dynamics of
perturbations can provide useful insights for understanding the origin of
adversarial sensitivities and developing robust solutions.
| stat |
Covariate powered cross-weighted multiple testing | A fundamental task in the analysis of datasets with many variables is
screening for associations. This can be cast as a multiple testing task, where
the objective is achieving high detection power while controlling type I error.
We consider $m$ hypothesis tests represented by pairs $((P_i, X_i))_{1\leq i
\leq m}$ of p-values $P_i$ and covariates $X_i$, such that $P_i \perp X_i$ if
$H_i$ is null. Here, we show how to use information potentially available in
the covariates about heterogeneities among hypotheses to increase power
compared to conventional procedures that only use the $P_i$. To this end, we
upgrade existing weighted multiple testing procedures through the Independent
Hypothesis Weighting (IHW) framework to use data-driven weights that are
calculated as a function of the covariates. Finite sample guarantees, e.g.,
false discovery rate (FDR) control, are derived from cross-weighting, a
data-splitting approach that enables learning the weight-covariate function
without overfitting as long as the hypotheses can be partitioned into
independent folds, with arbitrary within-fold dependence. IHW has increased
power compared to methods that do not use covariate information. A key
implication of IHW is that hypothesis rejection in common multiple testing
setups should not proceed according to the ranking of the p-values, but by an
alternative ranking implied by the covariate-weighted p-values.
| stat |
Hierarchical spline for time series forecasting: An application to Naval
ship engine failure rate | Predicting equipment failure is important because it could improve
availability and cut down the operating budget. Previous literature has
attempted to model failure rate with bathtub-formed function, Weibull
distribution, Bayesian network, or AHP. But these models perform well with a
sufficient amount of data and could not incorporate the two salient
characteristics; imbalanced category and sharing structure. Hierarchical model
has the advantage of partial pooling. The proposed model is based on Bayesian
hierarchical B-spline. Time series of the failure rate of 99 Republic of Korea
Naval ships are modeled hierarchically, where each layer corresponds to ship
engine, engine type, and engine archetype. As a result of the analysis, the
suggested model predicted the failure rate of an entire lifetime accurately in
multiple situational conditions, such as prior knowledge of the engine.
| stat |
A framework for automated anomaly detection in high frequency
water-quality data from in situ sensors | River water-quality monitoring is increasingly conducted using automated in
situ sensors, enabling timelier identification of unexpected values. However,
anomalies caused by technical issues confound these data, while the volume and
velocity of data prevent manual detection. We present a framework for automated
anomaly detection in high-frequency water-quality data from in situ sensors,
using turbidity, conductivity and river level data. After identifying end-user
needs and defining anomalies, we ranked their importance and selected suitable
detection methods. High priority anomalies included sudden isolated spikes and
level shifts, most of which were classified correctly by regression-based
methods such as autoregressive integrated moving average models. However, using
other water-quality variables as covariates reduced performance due to complex
relationships among variables. Classification of drift and periods of
anomalously low or high variability improved when we applied replaced anomalous
measurements with forecasts, but this inflated false positive rates.
Feature-based methods also performed well on high priority anomalies, but were
also less proficient at detecting lower priority anomalies, resulting in high
false negative rates. Unlike regression-based methods, all feature-based
methods produced low false positive rates, but did not and require training or
optimization. Rule-based methods successfully detected impossible values and
missing observations. Thus, we recommend using a combination of methods to
improve anomaly detection performance, whilst minimizing false detection rates.
Furthermore, our framework emphasizes the importance of communication between
end-users and analysts for optimal outcomes with respect to both detection
performance and end-user needs. Our framework is applicable to other types of
high frequency time-series data and anomaly detection applications.
| stat |
Maximum Likelihood Estimation and Graph Matching in Errorfully Observed
Networks | Given a pair of graphs with the same number of vertices, the inexact graph
matching problem consists in finding a correspondence between the vertices of
these graphs that minimizes the total number of induced edge disagreements. We
study this problem from a statistical framework in which one of the graphs is
an errorfully observed copy of the other. We introduce a corrupting channel
model, and show that in this model framework, the solution to the graph
matching problem is a maximum likelihood estimator. Necessary and sufficient
conditions for consistency of this MLE are presented, as well as a relaxed
notion of consistency in which a negligible fraction of the vertices need not
be matched correctly. The results are used to study matchability in several
families of random graphs, including edge independent models, random regular
graphs and small-world networks. We also use these results to introduce
measures of matching feasibility, and experimentally validate the results on
simulated and real-world networks.
| stat |
Preferential sampling for presence/absence data and for fusion of
presence/absence data with presence-only data | Presence/absence data and presence-only data are the two customary sources
for learning about species distributions over a region. We illuminate the
fundamental modeling differences between the two types of data. Most simply,
locations are considered as fixed under presence/absence data; locations are
random under presence-only data. The definition of "probability of presence" is
incompatible between the two. So, we take issue with modeling strategies in the
literature which ignore this incompatibility, which assume that
presence/absence modeling can be induced from presence-only specifications and
therefore, that fusion of presence-only and presence/absence data sources is
routine. We argue that presence/absence data should be modeled at point level.
That is, we need to specify a surface which provides the probability of
presence at any location in the region. A realization from this surface is a
binary map yielding the results of Bernoulli trials across all locations.
Presence-only data should be modeled as a point pattern driven by specification
of an intensity function. We further argue that, with just presence/absence
data, preferential sampling, using a shared process perspective, can improve
our estimated presence/absence surface and prediction of presence. We also
argue that preferential sampling can enable a probabilistically coherent fusion
of the two data types. We illustrate with two real datasets, one
presence/absence, one presence-only for invasive species presence in New
England in the United States. We demonstrate that potential bias in sampling
locations can affect inference with regard to presence/absence and show that
inference can be improved with preferential sampling ideas. We also provide a
probabilistically coherent fusion of the two datasets to again improve
inference with regard to presence/absence.
| stat |
Specification Tests for the Propensity Score | This paper proposes new nonparametric diagnostic tools to assess the
asymptotic validity of different treatment effects estimators that rely on the
correct specification of the propensity score. We derive a particular
restriction relating the propensity score distribution of treated and control
groups, and develop specification tests based upon it. The resulting tests do
not suffer from the "curse of dimensionality" when the vector of covariates is
high-dimensional, are fully data-driven, do not require tuning parameters such
as bandwidths, and are able to detect a broad class of local alternatives
converging to the null at the parametric rate $n^{-1/2}$, with $n$ the sample
size. We show that the use of an orthogonal projection on the tangent space of
nuisance parameters facilitates the simulation of critical values by means of a
multiplier bootstrap procedure, and can lead to power gains. The finite sample
performance of the tests is examined by means of a Monte Carlo experiment and
an empirical application. Open-source software is available for implementing
the proposed tests.
| stat |
Privacy for Spatial Point Process Data | In this work we develop methods for privatizing spatial location data, such
as spatial locations of individual disease cases. We propose two novel Bayesian
methods for generating synthetic location data based on log-Gaussian Cox
processes (LGCPs). We show that conditional predictive ordinate (CPO) estimates
can easily be obtained for point process data. We construct a novel risk metric
that utilizes CPO estimates to evaluate individual disclosure risks. We adapt
the propensity mean square error (pMSE) data utility metric for LGCPs. We
demonstrate that our synthesis methods offer an improved risk vs. utility
balance in comparison to radial synthesis with a case study of Dr. John Snow's
cholera outbreak data.
| stat |
Propensity score weighting under limited overlap and model
misspecification | Propensity score (PS) weighting methods are often used in non-randomized
studies to adjust for confounding and assess treatment effects. The most
popular among them, the inverse probability weighting (IPW), assigns weights
that are proportional to the inverse of the conditional probability of a
specific treatment assignment, given observed covariates. A key requirement for
IPW estimation is the positivity assumption, i.e., the PS must be bounded away
from 0 and 1. In practice, violations of the positivity assumption often
manifest by the presence of limited overlap in the PS distributions between
treatment groups. When these practical violations occur, a small number of
highly influential IPW weights may lead to unstable IPW estimators, with biased
estimates and large variances. To mitigate these issues, a number of
alternative methods have been proposed, including IPW trimming, overlap weights
(OW), matching weights (MW), and entropy weights (EW). Because OW, MW, and EW
target the population for whom there is equipoise (and with adequate overlap)
and their estimands depend on the true PS, a common criticism is that these
estimators may be more sensitive to misspecifications of the PS model. In this
paper, we conduct extensive simulation studies to compare the performances of
IPW and IPW trimming against those of OW, MW, and EW under limited overlap and
misspecified propensity score models. Across the wide range of scenarios we
considered, OW, MW, and EW consistently outperform IPW in terms of bias, root
mean squared error, and coverage probability.
| stat |
Joint Quantile Regression for Spatial Data | Linear quantile regression is a powerful tool to investigate how predictors
may affect a response heterogeneously across different quantile levels.
Unfortunately, existing approaches find it extremely difficult to adjust for
any dependency between observation units, largely because such methods are not
based upon a fully generative model of the data. For analyzing spatially
indexed data, we address this difficulty by generalizing the joint quantile
regression model of Yang and Tokdar (2017) and characterizing spatial
dependence via a Gaussian or $t$ copula process on the underlying quantile
levels of the observation units. A Bayesian semiparametric approach is
introduced to perform inference of model parameters and carry out spatial
quantile smoothing. An effective model comparison criteria is provided,
particularly for selecting between different model specifications of tail
heaviness and tail dependence. Extensive simulation studies and an application
to particulate matter concentration in northeast US are presented to illustrate
substantial gains in inference quality, accuracy and uncertainty quantification
over existing alternatives.
| stat |
Performance metrics for intervention-triggering prediction models do not
reflect an expected reduction in outcomes from using the model | Clinical researchers often select among and evaluate risk prediction models
using standard machine learning metrics based on confusion matrices. However,
if these models are used to allocate interventions to patients, standard
metrics calculated from retrospective data are only related to model utility
(in terms of reductions in outcomes) under certain assumptions. When
predictions are delivered repeatedly throughout time (e.g. in a patient
encounter), the relationship between standard metrics and utility is further
complicated. Several kinds of evaluations have been used in the literature, but
it has not been clear what the target of estimation is in each evaluation. We
synthesize these approaches, determine what is being estimated in each of them,
and discuss under what assumptions those estimates are valid. We demonstrate
our insights using simulated data as well as real data used in the design of an
early warning system. Our theoretical and empirical results show that
evaluations without interventional data either do not estimate meaningful
quantities, require strong assumptions, or are limited to estimating best-case
scenario bounds.
| stat |
Tractable Bayes of Skew-Elliptical Link Models for Correlated Binary
Data | Correlated binary response data with covariates are ubiquitous in
longitudinal or spatial studies. Among the existing statistical models the most
well-known one for this type of data is the multivariate probit model, which
uses a Gaussian link to model dependence at the latent level. However, a
symmetric link may not be appropriate if the data are highly imbalanced. Here,
we propose a multivariate skew-elliptical link model for correlated binary
responses, which includes the multivariate probit model as a special case.
Furthermore, we perform Bayesian inference for this new model and prove that
the regression coefficients have a closed-form unified skew-elliptical
posterior. The new methodology is illustrated by application to COVID-19
pandemic data from three different counties of the state of California, USA. By
jointly modeling extreme spikes in weekly new cases, our results show that the
spatial dependence cannot be neglected. Furthermore, the results also show that
the skewed latent structure of our proposed model improves the flexibility of
the multivariate probit model and provides better fit to our highly imbalanced
dataset.
| stat |
Factorisable Multitask Quantile Regression | A multivariate quantile regression model with a factor structure is proposed
to study data with many responses of interest. The factor structure is allowed
to vary with the quantile levels, which makes our framework more flexible than
the classical factor models. The model is estimated with the nuclear norm
regularization in order to accommodate the high dimensionality of data, but the
incurred optimization problem can only be efficiently solved in an approximate
manner by off-the-shelf optimization methods. Such a scenario is often seen
when the empirical risk is non-smooth or the numerical procedure involves
expensive subroutines such as singular value decomposition. To ensure that the
approximate estimator accurately estimates the model, non-asymptotic bounds on
error of the the approximate estimator is established. For implementation, a
numerical procedure that provably marginalizes the approximate error is
proposed. The merits of our model and the proposed numerical procedures are
demonstrated through Monte Carlo experiments and an application to finance
involving a large pool of asset returns.
| stat |
Fr\'echet random forests for metric space valued regression with non
euclidean predictors | Random forests are a statistical learning method widely used in many areas of
scientific research because of its ability to learn complex relationships
between input and output variables and also their capacity to handle
high-dimensional data. However, current random forest approaches are not
flexible enough to handle heterogeneous data such as curves, images and shapes.
In this paper, we introduce Fr\'echet trees and Fr\'echet random forests, which
allow to handle data for which input and output variables take values in
general metric spaces (which can be unordered). To this end, a new way of
splitting the nodes of trees is introduced and the prediction procedures of
trees and forests are generalized. Then, random forests out-of-bag error and
variable importance score are naturally adapted. A consistency theorem for
Fr\'echet regressogram predictor using data-driven partitions is given and
applied to Fr\'echet purely uniformly random trees. The method is studied
through several simulation scenarios on heterogeneous data combining
longitudinal, image and scalar data. Finally, two real datasets from HIV
vaccine trials are analyzed with the proposed method.
| stat |
High-dimensional structure learning of sparse vector autoregressive
models using fractional marginal pseudo-likelihood | Learning vector autoregressive models from multivariate time series is
conventionally approached through least squares or maximum likelihood
estimation. These methods typically assume a fully connected model which
provides no direct insight to the model structure and may lead to highly noisy
estimates of the parameters. Because of these limitations, there has been an
increasing interest towards methods that produce sparse estimates through
penalized regression. However, such methods are computationally intensive and
may become prohibitively time-consuming when the number of variables in the
model increases. In this paper we adopt an approximate Bayesian approach to the
learning problem by combining fractional marginal likelihood and
pseudo-likelihood. We propose a novel method, PLVAR, that is both faster and
produces more accurate estimates than the state-of-the-art methods based on
penalized regression. We prove the consistency of the PLVAR estimator and
demonstrate the attractive performance of the method on both simulated and
real-world data.
| stat |
Robustness of Model Predictions under Extension | Often, mathematical models of the real world are simplified representations
of complex systems. A caveat to using models for analysis is that predicted
causal effects and conditional independences may not be robust under model
extensions, and therefore applicability of such models is limited. In this
work, we consider conditions under which qualitative model predictions are
preserved when two models are combined. We show how to use the technique of
causal ordering to efficiently assess the robustness of qualitative model
predictions and characterize a large class of model extensions that preserve
these predictions. For dynamical systems at equilibrium, we demonstrate how
novel insights help to select appropriate model extensions and to reason about
the presence of feedback loops. We apply our ideas to a viral infection model
with immune responses.
| stat |
MetAL: Active Semi-Supervised Learning on Graphs via Meta Learning | The objective of active learning (AL) is to train classification models with
less number of labeled instances by selecting only the most informative
instances for labeling. The AL algorithms designed for other data types such as
images and text do not perform well on graph-structured data. Although a few
heuristics-based AL algorithms have been proposed for graphs, a principled
approach is lacking. In this paper, we propose MetAL, an AL approach that
selects unlabeled instances that directly improve the future performance of a
classification model. For a semi-supervised learning problem, we formulate the
AL task as a bilevel optimization problem. Based on recent work in
meta-learning, we use the meta-gradients to approximate the impact of
retraining the model with any unlabeled instance on the model performance.
Using multiple graph datasets belonging to different domains, we demonstrate
that MetAL efficiently outperforms existing state-of-the-art AL algorithms.
| stat |
Efficient sampling of conditioned Markov jump processes | We consider the task of generating draws from a Markov jump process (MJP)
between two time-points at which the process is known. Resulting draws are
typically termed bridges and the generation of such bridges plays a key role in
simulation-based inference algorithms for MJPs. The problem is challenging due
to the intractability of the conditioned process, necessitating the use of
computationally intensive methods such as weighted resampling or Markov chain
Monte Carlo. An efficient implementation of such schemes requires an
approximation of the intractable conditioned hazard/propensity function that is
both cheap and accurate. In this paper, we review some existing approaches to
this problem before outlining our novel contribution. Essentially, we leverage
the tractability of a Gaussian approximation of the MJP and suggest a
computationally efficient implementation of the resulting conditioned hazard
approximation. We compare and contrast our approach with existing methods using
three examples.
| stat |
Multiple Changepoint Detection with Partial Information on Changepoint
Times | This paper proposes a new minimum description length procedure to detect
multiple changepoints in time series data when some times are a priori thought
more likely to be changepoints. This scenario arises with temperature time
series homogenization pursuits, our focus here. Our Bayesian procedure
constructs a natural prior distribution for the situation, and is shown to
estimate the changepoint locations consistently, with an optimal convergence
rate. Our methods substantially improve changepoint detection power when prior
information is available. The methods are also tailored to bivariate data,
allowing changes to occur in one or both component series.
| stat |
Time-Varying Gaussian-Cauchy Mixture Models for Financial Risk
Management | There are various metrics for financial risk, such as value at risk (VaR),
expected shortfall, expected/unexpected loss, etc. When estimating these
metrics, it was very common to assume Gaussian distribution for the asset
returns, which may underestimate the real risk of the market, especially during
the financial crisis. In this paper, we propose a series of time-varying
mixture models for risk analysis and management. These mixture models contain
two components: one component with Gaussian distribution, and the other one
with a fat-tailed Cauchy distribution. We allow the distribution parameters and
component weights to change over time to increase the flexibility of the
models. Monte Carlo Expectation-Maximization algorithm is utilized to estimate
the parameters. To verify the good performance of our models, we conduct some
simulation studies, and implement our models to the real stock market. Based on
these studies, our models are appropriate under different economic conditions,
and the component weights can capture the correct pattern of the market
volatility.
| stat |
Linear Bandits with Stochastic Delayed Feedback | Stochastic linear bandits are a natural and well-studied model for structured
exploration/exploitation problems and are widely used in applications such as
online marketing and recommendation. One of the main challenges faced by
practitioners hoping to apply existing algorithms is that usually the feedback
is randomly delayed and delays are only partially observable. For example,
while a purchase is usually observable some time after the display, the
decision of not buying is never explicitly sent to the system. In other words,
the learner only observes delayed positive events. We formalize this problem as
a novel stochastic delayed linear bandit and propose ${\tt OTFLinUCB}$ and
${\tt OTFLinTS}$, two computationally efficient algorithms able to integrate
new information as it becomes available and to deal with the permanently
censored feedback. We prove optimal $\tilde O(\smash{d\sqrt{T}})$ bounds on the
regret of the first algorithm and study the dependency on delay-dependent
parameters. Our model, assumptions and results are validated by experiments on
simulated and real data.
| stat |
Grouped Variable Selection with Discrete Optimization: Computational and
Statistical Perspectives | We present a new algorithmic framework for grouped variable selection that is
based on discrete mathematical optimization. While there exist several
appealing approaches based on convex relaxations and nonconvex heuristics, we
focus on optimal solutions for the $\ell_0$-regularized formulation, a problem
that is relatively unexplored due to computational challenges. Our methodology
covers both high-dimensional linear regression and nonparametric sparse
additive modeling with smooth components. Our algorithmic framework consists of
approximate and exact algorithms. The approximate algorithms are based on
coordinate descent and local search, with runtimes comparable to popular sparse
learning algorithms. Our exact algorithm is based on a standalone
branch-and-bound (BnB) framework, which can solve the associated mixed integer
programming (MIP) problem to certified optimality. By exploiting the problem
structure, our custom BnB algorithm can solve to optimality problem instances
with $5 \times 10^6$ features in minutes to hours -- over $1000$ times larger
than what is currently possible using state-of-the-art commercial MIP solvers.
We also explore statistical properties of the $\ell_0$-based estimators. We
demonstrate, theoretically and empirically, that our proposed estimators have
an edge over popular group-sparse estimators in terms of statistical
performance in various regimes.
| stat |
MALTS: Matching After Learning to Stretch | We introduce a flexible framework that produces high-quality almost-exact
matches for causal inference. Most prior work in matching uses ad-hoc distance
metrics, often leading to poor quality matches, particularly when there are
irrelevant covariates. In this work, we learn an interpretable distance metric
for matching, which leads to substantially higher quality matches. The learned
distance metric stretches the covariates according to their contribution to
outcome prediction. The framework is flexible in that the user can choose the
form of the distance metric and the type of optimization algorithm. Our ability
to learn flexible distance metrics leads to matches that are interpretable and
useful for the estimation of conditional average treatment effects.
| stat |
Null Models and Community Detection in Multi-Layer Networks | Multi-layer networks are networks on a set of entities (nodes) with multiple
types of relations (edges) among them where each type of relation/interaction
is represented as a network layer. As with single layer networks, community
detection is an important task in multi-layer networks. A large group of
popular community detection methods in networks are based on optimizing a
quality function known as the modularity score, which is a measure of presence
of modules or communities in networks. Hence a first step in community
detection is defining a suitable modularity score that is appropriate for the
network in question. Here we introduce several multi-layer network modularity
measures under different null models of the network, motivated by empirical
observations in networks from a diverse field of applications. In particular we
define the multi-layer configuration model, the multi-layer expected degree
model and their various modifications as null models for multi-layer networks
to derive different modularities. The proposed modularities are grouped into
two categories. The first category, which is based on degree corrected
multi-layer stochastic block model, has the multi-layer expected degree model
as their null model. The second category, which is based on multi-layer
extensions of Newman-Girvan modularity, has the multi-layer configuration model
as their null model. These measures are then optimized to detect the optimal
community assignment of nodes. We compare the effectiveness of the measures in
community detection in simulated networks and then apply them to four real
networks.
| stat |
Bayesian wavelet de-noising with the caravan prior | According to both domain expert knowledge and empirical evidence, wavelet
coefficients of real signals tend to exhibit clustering patterns, in that they
contain connected regions of coefficients of similar magnitude (large or
small). A wavelet de-noising approach that takes into account such a feature of
the signal may in practice outperform other, more vanilla methods, both in
terms of the estimation error and visual appearance of the estimates. Motivated
by this observation, we present a Bayesian approach to wavelet de-noising,
where dependencies between neighbouring wavelet coefficients are a priori
modelled via a Markov chain-based prior, that we term the caravan prior.
Posterior computations in our method are performed via the Gibbs sampler. Using
representative synthetic and real data examples, we conduct a detailed
comparison of our approach with a benchmark empirical Bayes de-noising method
(due to Johnstone and Silverman). We show that the caravan prior fares well and
is therefore a useful addition to the wavelet de-noising toolbox.
| stat |
SPlit: An Optimal Method for Data Splitting | In this article we propose an optimal method referred to as SPlit for
splitting a dataset into training and testing sets. SPlit is based on the
method of Support Points (SP), which was initially developed for finding the
optimal representative points of a continuous distribution. We adapt SP for
subsampling from a dataset using a sequential nearest neighbor algorithm. We
also extend SP to deal with categorical variables so that SPlit can be applied
to both regression and classification problems. The implementation of SPlit on
real datasets shows substantial improvement in the worst-case testing
performance for several modeling methods compared to the commonly used random
splitting procedure.
| stat |
Variational inequalities and mean-field approximations for partially
observed systems of queueing networks | Queueing networks are systems of theoretical interest that find widespread
use in the performance evaluation of interconnected resources. In comparison to
counterpart models in genetics or mathematical biology, the stochastic (jump)
processes induced by queueing networks have distinctive coupling and
synchronization properties. This has prevented the derivation of variational
approximations for conditional representations of transient dynamics, which
rely on simplifying independence assumptions. Here, we present a model
augmentation to a multivariate counting process for interactions across service
stations, and we enable the variational evaluation of mean-field measures for
partially-observed multi-class networks. We also show that our framework offers
an efficient and improved alternative for inference tasks, where existing
variational or numerically intensive solutions do not work.
| stat |
Defence Against the Modern Arts: the Curse of Statistics "Score-based
likelihood ratios" | For several decades, legal and scientific scholars have argued that
conclusions from forensic examinations should be supported by statistical data
and reported within a probabilistic framework. Multiple models have been
proposed to quantify the probative value of forensic evidence. Unfortunately,
several of these models rely on ad-hoc strategies that are not scientifically
sound. The opacity of the technical jargon used to present these models and
their results, and the complexity of the techniques involved make it very
difficult for the untrained user to separate the wheat from the chaff. This
series of papers is intended to help forensic scientists and lawyers recognise
limitations and issues in tools proposed to interpret the results of forensic
examinations. This paper focuses on tools that have been proposed to leverage
the use of similarity scores to assess the probative value of forensic
findings. We call this family of tools "score-based likelihood ratios". In this
paper, we present the fundamental concepts on which these tools are built, we
describe some specific members of this family of tools, and we explore their
convergence to the Bayes factor through an intuitive geometrical approach and
through simulations. Finally, we discuss their validation and their potential
usefulness as a decision-making tool in forensic science.
| stat |
System modeling of a health issue: the case of preterm birth in Ohio | Preterm birth rate (PBR) stands out as a major public health concern in the
U.S. However, effective policies for mitigating the problem is largely unknown.
The complexities of the problem raise critical questions: Why is PBR increasing
despite the massive investment for reducing it? What policies can decrease it?
To address these questions, we develop a causal loop diagram to investigate
mechanisms underlying high preterm rate in a community. Our boundary is broad
and includes medical and education systems, as well as living conditions such
as crime rate and housing price. Then, we built a simulation model and divided
the population into two groups based on their chance of delivering a preterm
baby. We calibrated the model using the historical data of a case study,
Cuyahoga Ohio, from 1995 to 2017. Prior studies mostly applied reductionist
approaches to determine factors associated with high preterm rate at the
individual level. Our simulation model examines the reciprocal influences of
multiple factors and investigates the effect of different resource allocation
scenarios on the PBR. Results show that, in the case of Cuyahoga county with
one of the highest rates of PBR in the U.S., estimated preterm birth rates will
not be lower than the rates of 1995 during the next five years.
| stat |
Spectral estimation for spatial point patterns | This article determines how to implement spatial spectral analysis of point
processes (in two dimensions or more), by establishing the moments of raw
spectral summaries of point processes. We establish the first moments of raw
direct spectral estimates such as the discrete Fourier transform of a point
pattern. These have a number of surprising features that departs from the
properties of raw spectral estimates of random fields and time series. As for
random fields, the special case of isotropic processes warrants special
attention, which we discuss. For time series and random fields white noise
plays a special role, mirrored by the Poisson processes in the case of the
point process. For random fields bilinear estimators are prevalent in spectral
analysis. We discuss how to smooth any bilinear spectral estimator for a point
process. We also determine how to taper this bilinear spectral estimator, how
to calculate the periodogram, sample the wavenumbers and discuss the
correlation of the periodogram. In parts this corresponds to recommending
suitable separable as well as isotropic tapers in d dimensions. This, in
aggregation, establishes the foundations for spectral analysis of point
processes.
| stat |
Three-Dimensional Swarming Using Cyclic Stochastic Optimization | In this paper we simulate an ensemble of cooperating, mobile sensing agents
that implement the cyclic stochastic optimization (CSO) algorithm in an attempt
to survey and track multiple targets. In the CSO algorithm proposed, each agent
uses its sensed measurements, its shared information, and its predictions of
others' future motion to decide on its next action. This decision is selected
to minimize a loss function that decreases as the uncertainty in the targets'
state estimates decreases. Only noisy measurements of this loss function are
available to each agent, and in this study, each agent attempts to minimize
this function by calculating its stochastic gradient. This paper examines, via
simulation-based experiments, the implications and applicability of CSO
convergence in three dimensions.
| stat |
Meta-analysis of dichotomous and polytomous diagnostic tests without a
gold standard | Standard methods for the meta-analysis of diagnostic tests without a gold
standard are limited to dichotomous data. Multivariate probit models are used
to analyze correlated binary data, and can be extended to multivariate ordered
probit models to model polytomous (i.e. non-binary) data. Within the context of
an imperfect gold standard, they have previously been used for the analysis of
dichotomous and polytomous diagnostic tests in a single study, and for the
meta-analysis of dichotomous tests. In this paper, we developed a hierarchical,
latent class multivariate probit model for the meta-analysis of polytomous and
dichotomous diagnostic tests without a gold standard. The model can accommodate
a hierarchical partial pooling model on the conditional within-study
correlations, which allow us to obtain summary estimates of joint test
accuracy. Dichotomous tests use probit regression likelihoods and polytomous
tests use ordinal probit regression likelihoods. We fitted the models using
Stan, which uses a state-of-the-art Hamiltonian Monte Carlo algorithm. We
applied the models to a dataset in which studies evaluated the accuracy of
tests, and combinations of tests, for deep vein thrombosis. We first
demonstrate the issues with dichotomising test accuracy data a priori without a
gold standard, and then we apply a model which does not dichotomise the data.
We fitted models assuming conditional independence and dependence between
tests, as well as models assuming a perfect gold standard, and compared model
fit and summary estimates.
| stat |
A generalization of the symmetrical and optimal
probability-to-possibility transformations | Possibility and probability theories are alternative and complementary ways
to deal with uncertainty, which has motivated over the last years an interest
for the study of ways to transform probability distributions into possibility
distributions and conversely. This paper studies the advantages and
shortcomings of two well-known discrete probability to possibility
transformations: the optimal transformation and the symmetrical transformation,
and presents a novel parametric family of probability to possibility
transformations which generalizes them and alleviate their shortcomings,
showing a big potential for practical application. The paper also introduces a
novel fuzzy measure of specificity for probability distributions based on the
concept of fuzzy subsethood and presents a empirical validation of the
generalized transformation usefulness applying it to the text authorship
attribution problem.
| stat |
Rotting bandits are not harder than stochastic ones | In stochastic multi-armed bandits, the reward distribution of each arm is
assumed to be stationary. This assumption is often violated in practice (e.g.,
in recommendation systems), where the reward of an arm may change whenever is
selected, i.e., rested bandit setting. In this paper, we consider the
non-parametric rotting bandit setting, where rewards can only decrease. We
introduce the filtering on expanding window average (FEWA) algorithm that
constructs moving averages of increasing windows to identify arms that are more
likely to return high rewards when pulled once more. We prove that for an
unknown horizon $T$, and without any knowledge on the decreasing behavior of
the $K$ arms, FEWA achieves problem-dependent regret bound of
$\widetilde{\mathcal{O}}(\log{(KT)}),$ and a problem-independent one of
$\widetilde{\mathcal{O}}(\sqrt{KT})$. Our result substantially improves over
the algorithm of Levine et al. (2017), which suffers regret
$\widetilde{\mathcal{O}}(K^{1/3}T^{2/3})$. FEWA also matches known bounds for
the stochastic bandit setting, thus showing that the rotting bandits are not
harder. Finally, we report simulations confirming the theoretical improvements
of FEWA.
| stat |
A Simple Approach to Online Sparse Sliced Inverse Regression | Sliced inverse regression is an efficient approach to estimate the central
subspace for sufficient dimension reduction. Due to the demand for tackling the
problem of sparse high dimensional data, several methods of online sufficient
dimension reduction has been proposed. However, as far as we know, all of these
methods are not well suitable for high dimensional and sparse data. Hence, the
purpose of this paper is to propose a simple and efficient approach to online
sparse sliced inverse regression (OSSIR). Motivated by Lasso-SIR and online
SIR, we implement the Lasso-SIR in an online fashion. There are two important
steps in our method, one is to iteratively obtain the eigenvalues and
eigenvectors of matrix $\operatorname{cov}(E(x|Y))$, the other is the online
$L_{1}$ regularization. For the former problem, we expand the online principal
component analysis and summarize four different ways. While in the online
fashion, truncated gradient has been shown to be an online counterpart of
$L_{1}$ regularization in the batch setting, so we apply the truncated gradient
in the online sliced inverse regression for the latter problem. The theoretical
properties of this online learner are established. By comparing with several
existing methods in the simulations and real data applications, we demonstrate
the effectiveness and efficiency of our algorithm.
| stat |
Constraining Logits by Bounded Function for Adversarial Robustness | We propose a method for improving adversarial robustness by addition of a new
bounded function just before softmax. Recent studies hypothesize that small
logits (inputs of softmax) by logit regularization can improve adversarial
robustness of deep learning. Following this hypothesis, we analyze norms of
logit vectors at the optimal point under the assumption of universal
approximation and explore new methods for constraining logits by addition of a
bounded function before softmax. We theoretically and empirically reveal that
small logits by addition of a common activation function, e.g., hyperbolic
tangent, do not improve adversarial robustness since input vectors of the
function (pre-logit vectors) can have large norms. From the theoretical
findings, we develop the new bounded function. The addition of our function
improves adversarial robustness because it makes logit and pre-logit vectors
have small norms. Since our method only adds one activation function before
softmax, it is easy to combine our method with adversarial training. Our
experiments demonstrate that our method is comparable to logit regularization
methods in terms of accuracies on adversarially perturbed datasets without
adversarial training. Furthermore, it is superior or comparable to logit
regularization methods and a recent defense method (TRADES) when using
adversarial training.
| stat |
Conditional canonical correlation estimation based on covariates with
random forests | Investigating the relationships between two sets of variables helps to
understand their interactions and can be done with canonical correlation
analysis (CCA). However, the correlation between the two sets can sometimes
depend on a third set of covariates, often subject-related ones such as age,
gender, or other clinical measures. In this case, applying CCA to the whole
population is not optimal and methods to estimate conditional CCA, given the
covariates, can be useful. We propose a new method called Random Forest with
Canonical Correlation Analysis (RFCCA) to estimate the conditional canonical
correlations between two sets of variables given subject-related covariates.
The individual trees in the forest are built with a splitting rule specifically
designed to partition the data to maximize the canonical correlation
heterogeneity between child nodes. We also propose a significance test to
detect the global effect of the covariates on the relationship between two sets
of variables. The performance of the proposed method and the global
significance test is evaluated through simulation studies that show it provides
accurate canonical correlation estimations and well-controlled Type-1 error. We
also show an application of the proposed method with EEG data.
| stat |
Confidence Intervals for Nonparametric Empirical Bayes Analysis | In an empirical Bayes analysis, we use data from repeated sampling to imitate
inferences made by an oracle Bayesian with extensive knowledge of the
data-generating distribution. Existing results provide a comprehensive
characterization of when and why empirical Bayes point estimates accurately
recover oracle Bayes behavior. In this paper, we develop flexible and practical
confidence intervals that provide asymptotic frequentist coverage of empirical
Bayes estimands, such as the posterior mean or the local false sign rate. The
coverage statements hold even when the estimands are only partially identified
or when empirical Bayes point estimates converge very slowly.
| stat |
Network modelling of topological domains using Hi-C data | Chromosome conformation capture experiments such as Hi-C are used to map the
three-dimensional spatial organization of genomes. One specific feature of the
3D organization is known as topologically associating domains (TADs), which are
densely interacting, contiguous chromatin regions playing important roles in
regulating gene expression. A few algorithms have been proposed to detect TADs.
In particular, the structure of Hi-C data naturally inspires application of
community detection methods. However, one of the drawbacks of community
detection is that most methods take exchangeability of the nodes in the network
for granted; whereas the nodes in this case, i.e. the positions on the
chromosomes, are not exchangeable. We propose a network model for detecting
TADs using Hi-C data that takes into account this non-exchangeability. In
addition, our model explicitly makes use of cell-type specific CTCF binding
sites as biological covariates and can be used to identify conserved TADs
across multiple cell types. The model leads to a likelihood objective that can
be efficiently optimized via relaxation. We also prove that when suitably
initialized, this model finds the underlying TAD structure with high
probability. Using simulated data, we show the advantages of our method and the
caveats of popular community detection methods, such as spectral clustering, in
this application. Applying our method to real Hi-C data, we demonstrate the
domains identified have desirable epigenetic features and compare them across
different cell types.
| stat |
Specification tests in semiparametric transformation models - a
multiplier bootstrap approach | We consider semiparametric transformation models, where after pre-estimation
of a parametric transformation of the response the data are modeled by means of
nonparametric regression. We suggest subsequent procedures for testing
lack-of-fit of the regression function and for significance of covariables,
which - in contrast to procedures from the literature - are asymptotically not
influenced by the pre-estimation of the transformation. The test statistics are
asymptotically pivotal and have the same asymptotic distribution as in
regression models without transformation. We show validity of a multiplier
bootstrap procedure which is easier to implement and much less computationally
demanding than bootstrap procedures based on the transformation model. In a
simulation study we demonstrate the superior performance of the procedure in
comparison with the competitors from the literature.
| stat |
Clustered active-subspace based local Gaussian Process emulator for
high-dimensional and complex computer models | Quantifying uncertainties in physical or engineering systems often requires a
large number of simulations of the underlying computer models that are
computationally intensive. Emulators or surrogate models are often used to
accelerate the computation in such problems, and in this regard the Gaussian
Process (GP) emulator is a popular choice for its ability to quantify the
approximation error in the emulator itself. However, a major limitation of the
GP emulator is that it can not handle problems of very high dimensions, which
is often addressed with dimension reduction techniques. In this work we hope to
address an issue that the models of interest are so complex that they admit
different low dimensional structures in different parameter regimes. Building
upon the active subspace method for dimension reduction, we propose a clustered
active subspace method which identifies the local low-dimensional structures as
well as the parameter regimes they are in (represented as clusters), and then
construct low dimensional and local GP emulators within the clusters.
Specifically we design a clustering method based on the gradient information to
identify these clusters, and a local GP construction procedure to construct the
GP emulator within a local cluster. With numerical examples, we demonstrate
that the proposed method is effective when the underlying models are of complex
low-dimensional structures.
| stat |
Adaptive Covariate Acquisition for Minimizing Total Cost of
Classification | In some applications, acquiring covariates comes at a cost which is not
negligible. For example in the medical domain, in order to classify whether a
patient has diabetes or not, measuring glucose tolerance can be expensive.
Assuming that the cost of each covariate, and the cost of misclassification can
be specified by the user, our goal is to minimize the (expected) total cost of
classification, i.e. the cost of misclassification plus the cost of the
acquired covariates. We formalize this optimization goal using the
(conditional) Bayes risk and describe the optimal solution using a recursive
procedure. Since the procedure is computationally infeasible, we consequently
introduce two assumptions: (1) the optimal classifier can be represented by a
generalized additive model, (2) the optimal sets of covariates are limited to a
sequence of sets of increasing size. We show that under these two assumptions,
a computationally efficient solution exists. Furthermore, on several medical
datasets, we show that the proposed method achieves in most situations the
lowest total costs when compared to various previous methods. Finally, we
weaken the requirement on the user to specify all misclassification costs by
allowing the user to specify the minimally acceptable recall (target recall).
Our experiments confirm that the proposed method achieves the target recall
while minimizing the false discovery rate and the covariate acquisition costs
better than previous methods.
| stat |
Median regression with differential privacy | Median regression analysis has robustness properties which make it attractive
compared with regression based on the mean, while differential privacy can
protect individual privacy during statistical analysis of certain datasets. In
this paper, three privacy preserving methods are proposed for median
regression. The first algorithm is based on a finite smoothing method, the
second provides an iterative way and the last one further employs the greedy
coordinate descent approach. Privacy preserving properties of these three
methods are all proved. Accuracy bound or convergence properties of these
algorithms are also provided. Numerical calculation shows that the first method
has better accuracy than the others when the sample size is small. When the
sample size becomes larger, the first method needs more time while the second
method needs less time with well-matched accuracy. For the third method, it
costs less time in both cases, while it highly depends on step size.
| stat |
Robust Function-on-Function Regression | Functional linear regression is a widely used approach to model functional
responses with respect to functional inputs. However, classical functional
linear regression models can be severely affected by outliers. We therefore
introduce a Fisher-consistent robust functional linear regression model that is
able to effectively fit data in the presence of outliers. The model is built
using robust functional principal component and least squares regression
estimators. The performance of the functional linear regression model depends
on the number of principal components used. We therefore introduce a consistent
robust model selection procedure to choose the number of principal components.
Our robust functional linear regression model can be used alongside an outlier
detection procedure to effectively identify abnormal functional responses. A
simulation study shows our method is able to effectively capture the regression
behaviour in the presence of outliers, and is able to find the outliers with
high accuracy. We demonstrate the usefulness of our method on jet engine sensor
data. We identify outliers that would not be found if the functional responses
were modelled independently of the functional input, or using non-robust
methods.
| stat |
Parallel subgroup analysis of high-dimensional data via M-regression | It becomes an interesting problem to identify subgroup structures in data
analysis as populations are probably heterogeneous in practice. In this paper,
we consider M-estimators together with both concave and pairwise fusion
penalties, which can deal with high-dimensional data containing some outliers.
The penalties are applied both on covariates and treatment effects, where the
estimation is expected to achieve both variable selection and data clustering
simultaneously. An algorithm is proposed to process relatively large datasets
based on parallel computing. We establish the convergence analysis of the
proposed algorithm, the oracle property of the penalized M-estimators, and the
selection consistency of the proposed criterion. Our numerical study
demonstrates that the proposed method is promising to efficiently identify
subgroups hidden in high-dimensional data.
| stat |
Ensemble Kalman Variational Objectives: Nonlinear Latent Trajectory
Inference with A Hybrid of Variational Inference and Ensemble Kalman Filter | Variational Inference (VI) combined with Bayesian nonlinear filtering
produces the state-of-the-art results for latent trajectory inference. A body
of recent works focused on Sequential Monte Carlo (SMC) and its expansion,
e.g., Forward Filtering Backward Simulation (FFBSi). These studies achieved a
great success, however, remain a serious problem for particle degeneracy. In
this paper, we propose Ensemble Kalman Objectives (EnKOs), the hybrid method of
VI and Ensemble Kalman Filter (EnKF), to infer the State Space Models (SSMs).
Unlike the SMC based methods, the our proposed method can identify the latent
dynamics given fewer particles because of its rich particle diversity. We
demonstrate that EnKOs outperform the SMC based methods in terms of predictive
ability for three benchmark nonlinear dynamics systems tasks.
| stat |
AVaN Pack: An Analytical/Numerical Solution for Variance-Based
Sensitivity Analysis | Sensitivity analysis is an important concept to analyze the influences of
parameters in a system, an equation or a collection of data. The methods used
for sensitivity analysis are divided into deterministic and statistical
techniques. Generally, deterministic techniques analyze fixed points of a model
whilst stochastic techniques analyze a range of values. Deterministic methods
fail in analyze the entire range of input values and stochastic methods
generate outcomes with random errors. In this manuscript, we are interested in
stochastic methods, mainly in variance-based techniques such as Variance and
Sobol indices, since this class of techniques is largely used on literature.
The objective of this manuscript is to present an analytical solution for
variance based sensitive analysis. As a result of this research, two small
programs were developed in Javascript named as AVaN Pack (Analysis of Variance
through Numerical solution). These programs allow users to find the
contribution of each individual parameter in any function by means of a
mathematical solution, instead of sampling-based ones.
| stat |
Minimum adjusted Rand index for two clusterings of a given size | The adjusted Rand index (ARI) is commonly used in cluster analysis to measure
the degree of agreement between two data partitions. Since its introduction,
exploring the situations of extreme agreement and disagreement under different
circumstances has been a subject of interest, in order to achieve a better
understanding of this index. Here, an explicit formula for the lowest possible
value of the ARI for two clusterings of given sizes is shown, and moreover a
specific pair of clusterings achieving such a bound is provided.
| stat |
Asymptotic distribution-free change-point detection for data with
repeated observations | In the regime of change-point detection, a nonparametric framework based on
scan statistics utilizing graphs representing similarities among observations
is gaining attention due to its flexibility and good performances for
high-dimensional and non-Euclidean data sequences, which are ubiquitous in this
big data era. However, this graph-based framework encounters problems when
there are repeated observations in the sequence, which often happens for
discrete data, such as network data. In this work, we extend the graph-based
framework to solve this problem by averaging or taking union of all possible
"optimal" graphs resulted from repeated observations. We consider both the
single change-point alternative and the changed-interval alternative, and
derive analytic formulas to control the type I error for the new methods,
making them fast applicable to large data sets. The extended methods are
illustrated on an application in detecting changes in a sequence of dynamic
networks over time.
| stat |
Robust Sparse Reduced Rank Regression in High Dimensions | We propose robust sparse reduced rank regression for analyzing large and
complex high-dimensional data with heavy-tailed random noise. The proposed
method is based on a convex relaxation of a rank- and sparsity-constrained
non-convex optimization problem, which is then solved using the alternating
direction method of multipliers algorithm. We establish non-asymptotic
estimation error bounds under both Frobenius and nuclear norms in the
high-dimensional setting. This is a major contribution over existing results in
reduced rank regression, which mainly focus on rank selection and prediction
consistency. Our theoretical results quantify the tradeoff between
heavy-tailedness of the random noise and statistical bias. For random noise
with bounded $(1+\delta)$th moment with $\delta \in (0,1)$, the rate of
convergence is a function of $\delta$, and is slower than the sub-Gaussian-type
deviation bounds; for random noise with bounded second moment, we obtain a rate
of convergence as if sub-Gaussian noise were assumed. Furthermore, the
transition between the two regimes is smooth. We illustrate the performance of
the proposed method via extensive numerical studies and a data application.
| stat |
Efficient Bayesian synthetic likelihood with whitening transformations | Likelihood-free methods are an established approach for performing
approximate Bayesian inference for models with intractable likelihood
functions. However, they can be computationally demanding. Bayesian synthetic
likelihood (BSL) is a popular such method that approximates the likelihood
function of the summary statistic with a known, tractable distribution --
typically Gaussian -- and then performs statistical inference using standard
likelihood-based techniques. However, as the number of summary statistics
grows, the number of model simulations required to accurately estimate the
covariance matrix for this likelihood rapidly increases. This poses significant
challenge for the application of BSL, especially in cases where model
simulation is expensive. In this article we propose whitening BSL (wBSL) -- an
efficient BSL method that uses approximate whitening transformations to
decorrelate the summary statistics at each algorithm iteration. We show
empirically that this can reduce the number of model simulations required to
implement BSL by more than an order of magnitude, without much loss of
accuracy. We explore a range of whitening procedures and demonstrate the
performance of wBSL on a range of simulated and real modelling scenarios from
ecology and biology.
| stat |
Online Particle Smoothing with Application to Map-matching | We introduce a novel method for online smoothing in state-space models based
on a fixed-lag approximation. Unlike classical fixed-lag smoothing we
approximate the joint posterior distribution rather than just the marginals. By
only partially resampling particles, our online particle smoothing technique
avoids path degeneracy as the length of the state-space model increases. We
demonstrate the utility of our method in the context of map-matching, the task
of inferring a vehicle's trajectory given a road network and noisy GPS
observations.
| stat |
Risk Bounds for Learning Multiple Components with Permutation-Invariant
Losses | This paper proposes a simple approach to derive efficient error bounds for
learning multiple components with sparsity-inducing regularization. We show
that for such regularization schemes, known decompositions of the Rademacher
complexity over the components can be used in a more efficient manner to result
in tighter bounds without too much effort. We give examples of application to
switching regression and center-based clustering/vector quantization. Then, the
complete workflow is illustrated on the problem of subspace clustering, for
which decomposition results were not previously available. For all these
problems, the proposed approach yields risk bounds with mild dependencies on
the number of components and completely removes this dependence for nonconvex
regularization schemes that could not be handled by previous methods.
| stat |
Distributionally Robust Bayesian Optimization | Robustness to distributional shift is one of the key challenges of
contemporary machine learning. Attaining such robustness is the goal of
distributionally robust optimization, which seeks a solution to an optimization
problem that is worst-case robust under a specified distributional shift of an
uncontrolled covariate. In this paper, we study such a problem when the
distributional shift is measured via the maximum mean discrepancy (MMD). For
the setting of zeroth-order, noisy optimization, we present a novel
distributionally robust Bayesian optimization algorithm (DRBO). Our algorithm
provably obtains sub-linear robust regret in various settings that differ in
how the uncertain covariate is observed. We demonstrate the robust performance
of our method on both synthetic and real-world benchmarks.
| stat |
VFlow: More Expressive Generative Flows with Variational Data
Augmentation | Generative flows are promising tractable models for density modeling that
define probabilistic distributions with invertible transformations. However,
tractability imposes architectural constraints on generative flows, making them
less expressive than other types of generative models. In this work, we study a
previously overlooked constraint that all the intermediate representations must
have the same dimensionality with the original data due to invertibility,
limiting the width of the network. We tackle this constraint by augmenting the
data with some extra dimensions and jointly learning a generative flow for
augmented data as well as the distribution of augmented dimensions under a
variational inference framework. Our approach, VFlow, is a generalization of
generative flows and therefore always performs better. Combining with existing
generative flows, VFlow achieves a new state-of-the-art 2.98 bits per dimension
on the CIFAR-10 dataset and is more compact than previous models to reach
similar modeling quality.
| stat |
Empirical Risk Minimization and Stochastic Gradient Descent for
Relational Data | Empirical risk minimization is the main tool for prediction problems, but its
extension to relational data remains unsolved. We solve this problem using
recent ideas from graph sampling theory to (i) define an empirical risk for
relational data and (ii) obtain stochastic gradients for this empirical risk
that are automatically unbiased. This is achieved by considering the method by
which data is sampled from a graph as an explicit component of model design. By
integrating fast implementations of graph sampling schemes with standard
automatic differentiation tools, we provide an efficient turnkey solver for the
risk minimization problem. We establish basic theoretical properties of the
procedure. Finally, we demonstrate relational ERM with application to two
non-standard problems: one-stage training for semi-supervised node
classification, and learning embedding vectors for vertex attributes.
Experiments confirm that the turnkey inference procedure is effective in
practice, and that the sampling scheme used for model specification has a
strong effect on model performance. Code is available at
https://github.com/wooden-spoon/relational-ERM.
| stat |
Selective Inference for Latent Block Models | Model selection in latent block models has been a challenging but important
task in the field of statistics. Specifically, a major challenge is encountered
when constructing a test on a block structure obtained by applying a specific
clustering algorithm to a finite size matrix. In this case, it becomes crucial
to consider the selective bias in the block structure, that is, the block
structure is selected from all the possible cluster memberships based on some
criterion by the clustering algorithm. To cope with this problem, this study
provides a selective inference method for latent block models. Specifically, we
construct a statistical test on a set of row and column cluster memberships of
a latent block model, which is given by a squared residue minimization
algorithm. The proposed test, by its nature, includes and thus can also be used
as the test on the set of row and column cluster numbers. We also propose an
approximated version of the test based on simulated annealing to avoid
combinatorial explosion in searching the optimal block structure. The results
show that the proposed exact and approximated tests work effectively, compared
to the naive test that did not take the selective bias into account.
| stat |
Simultaneous Confidence Tubes for Comparison of Several Multivariate
Linear Regression Models | Much of the research on multiple comparison and simultaneous inference in the
past sixty years or so has been for the comparisons of several population
means. Spurrier (1999) seems to be the first to study the multiple comparison
of several simple linear regression lines by using simultaneous confidence
bands. In this paper, the work of Liu et al. (2004) for finite comparisons of
several univariate linear regression models by using simultaneous confidence
bands has been extended to finite comparison of several multivariate linear
regression models by using simultaneous confidence tubes. We show how
simultaneous confidence tubes can be constructed to allow more informative
inferences for the comparison of several multivariate linear regression models
than the current approach of hypotheses testing. The methodologies are
illustrated with examples.
| stat |
Learning Sparse Nonparametric DAGs | We develop a framework for learning sparse nonparametric directed acyclic
graphs (DAGs) from data. Our approach is based on a recent algebraic
characterization of DAGs that led to a fully continuous program for score-based
learning of DAG models parametrized by a linear structural equation model
(SEM). We extend this algebraic characterization to nonparametric SEM by
leveraging nonparametric sparsity based on partial derivatives, resulting in a
continuous optimization problem that can be applied to a variety of
nonparametric and semiparametric models including GLMs, additive noise models,
and index models as special cases. Unlike existing approaches that require
specific modeling choices, loss functions, or algorithms, we present a
completely general framework that can be applied to general nonlinear models
(e.g. without additive noise), general differentiable loss functions, and
generic black-box optimization routines. The code is available at
https://github.com/xunzheng/notears.
| stat |
Is There a Trade-Off Between Fairness and Accuracy? A Perspective Using
Mismatched Hypothesis Testing | A trade-off between accuracy and fairness is almost taken as a given in the
existing literature on fairness in machine learning. Yet, it is not preordained
that accuracy should decrease with increased fairness. Novel to this work, we
examine fair classification through the lens of mismatched hypothesis testing:
trying to find a classifier that distinguishes between two ideal distributions
when given two mismatched distributions that are biased. Using Chernoff
information, a tool in information theory, we theoretically demonstrate that,
contrary to popular belief, there always exist ideal distributions such that
optimal fairness and accuracy (with respect to the ideal distributions) are
achieved simultaneously: there is no trade-off. Moreover, the same classifier
yields the lack of a trade-off with respect to ideal distributions while
yielding a trade-off when accuracy is measured with respect to the given
(possibly biased) dataset. To complement our main result, we formulate an
optimization to find ideal distributions and derive fundamental limits to
explain why a trade-off exists on the given biased dataset. We also derive
conditions under which active data collection can alleviate the
fairness-accuracy trade-off in the real world. Our results lead us to contend
that it is problematic to measure accuracy with respect to data that reflects
bias, and instead, we should be considering accuracy with respect to ideal,
unbiased data.
| stat |
Estimation of local treatment under the binary instrumental variable
model | Instrumental variables are widely used to deal with unmeasured confounding in
observational studies and imperfect randomized controlled trials. In these
studies, researchers often target the so-called local average treatment effect
as it is identifiable under mild conditions. In this paper, we consider
estimation of the local average treatment effect under the binary instrumental
variable model. We discuss the challenges for causal estimation with a binary
outcome, and show that surprisingly, it can be more difficult than the case
with a continuous outcome. We propose novel modeling and estimating procedures
that improve upon existing proposals in terms of model congeniality,
interpretability, robustness or efficiency. Our approach is illustrated via
simulation studies and a real data analysis.
| stat |
An Extreme Value Bayesian Lasso for the Conditional Bulk and Tail | We introduce a novel regression model for the conditional bulk and
conditional tail of a possibly heavy-tailed response. The proposed model can be
used to learn the effect of covariates on an extreme value setting via a
Lasso-type specification based on a Lagrangian restriction. Our model can be
used to track if some covariates are significant for the bulk, but not for the
tail---and vice-versa; in addition to this, the proposed model bypasses the
need for conditional threshold selection in an extreme value theory framework.
We assess the finite-sample performance of the proposed methods through a
simulation study that reveals that our method recovers the true conditional
distribution over a variety of simulation scenarios, along with being accurate
on variable selection. Rainfall data are used to showcase how the proposed
method can learn to distinguish between key drivers of moderate rainfall,
against those of extreme rainfall.
| stat |
Sensitivity Analysis of Error-Contaminated Time Series Data under
Autoregressive Models with Application of COVID-19 Data | Autoregressive (AR) models are useful tools in time series analysis.
Inferences under such models are distorted in the presence of measurement
error, which is very common in practice. In this article, we establish
analytical results for quantifying the biases of the parameter estimation in AR
models if the measurement error effects are neglected. We propose two
measurement error models to describe different processes of data contamination.
An estimating equation approach is proposed for the estimation of the model
parameters with measurement error effects accounted for. We further discuss
forecasting using the proposed method. Our work is inspired by COVID-19 data,
which are error-contaminated due to multiple reasons including the asymptomatic
cases and varying incubation periods. We implement our proposed method by
conducting sensitivity analyses and forecasting of the mortality rate of
COVID-19 over time for the four most populated provinces in Canada. The results
suggest that incorporating or not incorporating measurement error effects
yields rather different results for parameter estimation and forecasting.
| stat |
Generalized Elliptical Slice Sampling with Regional Pseudo-priors | In this paper, we propose a MCMC algorithm based on elliptical slice sampling
with the purpose to improve sampling efficiency. During sampling, a mixture
distribution is fitted periodically to previous samples. The components of the
mixture distribution are called regional pseudo-priors because each component
serves as the pseudo-prior for a subregion of the sampling space. Expectation
maximization algorithm, variational inference algorithm and stochastic
approximation algorithm are used to estimate the parameters. Meanwhile,
parallel computing is used to relieve the burden of computation. Ergodicity of
the proposed algorithm is proven mathematically. Experimental results on one
synthetic and two real-world dataset show that the proposed algorithm has the
following advantages: with the same starting points, the proposed algorithm can
find more distant modes; the proposed algorithm has lower rejection rates; when
doing Bayesian inference for uni-modal posterior distributions, the proposed
algorithm can give more accurate estimations; when doing Bayesian inference for
multi-modal posterior distributions, the proposed algorithm can find different
modes well, and the estimated means of the mixture distribution can provide
additional information for the location of modes.
| stat |
A survey on natural language processing (nlp) and applications in
insurance | Text is the most widely used means of communication today. This data is
abundant but nevertheless complex to exploit within algorithms. For years,
scientists have been trying to implement different techniques that enable
computers to replicate some mechanisms of human reading. During the past five
years, research disrupted the capacity of the algorithms to unleash the value
of text data. It brings today, many opportunities for the insurance
industry.Understanding those methods and, above all, knowing how to apply them
is a major challenge and key to unleash the value of text data that have been
stored for many years. Processing language with computer brings many new
opportunities especially in the insurance sector where reports are central in
the information used by insurers. SCOR's Data Analytics team has been working
on the implementation of innovative tools or products that enable the use of
the latest research on text analysis. Understanding text mining techniques in
insurance enhances the monitoring of the underwritten risks and many processes
that finally benefit policyholders.This article proposes to explain
opportunities that Natural Language Processing (NLP) are providing to
insurance. It details different methods used today in practice traces back the
story of them. We also illustrate the implementation of certain methods using
open source libraries and python codes that we have developed to facilitate the
use of these techniques.After giving a general overview on the evolution of
text mining during the past few years,we share about how to conduct a full
study with text mining and share some examples to serve those models into
insurance products or services. Finally, we explained in more details every
step that composes a Natural Language Processing study to ensure the reader can
have a deep understanding on the implementation.
| stat |
Predication of Inflection Point and Outbreak Size of COVID-19 in New
Epicentres | The coronavirus disease 2019 (COVID-19) had caused more that 8 million
infections as of middle June 2020. Recently, Brazil has become a new epicentre
of COVID-19, while India and African region are potential epicentres. This
study aims to predict the inflection point and outbreak size of these
new/potential epicentres at the early phase of the epidemics by borrowing
information from more `mature' curves from other countries. We modeled the
cumulative cases to the well-known sigmoid growth curves to describe the
epidemic trends under the mixed-effect models and using the four-parameter
logistic model after power transformations. African region is predicted to have
the largest total outbreak size of 3.9 million cases (2.2 to 6 million), and
the inflection will come around September 13, 2020. Brazil and India are
predicted to have a similar final outbreak size of around 2.5 million cases
(1.1 to 4.3 million), with the inflection points arriving June 23 and July 26,
respectively. We conclude in Brazil, India, and African the epidemics of COVI19
have not yet passed the inflection points; these regions potentially can take
over USA in terms of outbreak size
| stat |
Encouraging Equitable Bikeshare: Implications of Docked and Dockless
Models for Spatial Equity | The last decade has seen a rapid rise in the number of bikeshare programs,
where bikes are made available throughout a community on an as-needed basis.
Given that many of these programs are at least partially publicly funded, a
central concern of operators and investors is whether these systems operate
equitably. Though spatial equity has been well-studied under the docked model,
where bikes are picked up and dropped off at prespecified docking stations,
there has been little work examining that of the increasingly popular dockless
model, where bikes can be picked up and dropped off from anywhere within an
operating area. We explore comparative equity in spatial access to bikeshare
services under these two models by collecting spatial data on 45,935 bikes from
73 bikeshare systems using a novel querying approach (with generalizable and
freely available source code), and joining this data with newly-available
sociodemographic data at the census tract level. Using Poisson count
regression, we perform the first comparative analysis of the two docking
approaches, finding that dockless systems operate more equitably than docked
systems by education, but do not differ in spatial access by socioeconomic
class.
| stat |
Moderating effects of retail operations and hard-sell sales techniques
on salesperson's interpersonal skills and customer repurchase intention | Salesperson's interpersonal skills have always played an important role in
influencing various stages of customer's purchase decision. With the increase
in retail outlets and merchandisers, retail operations have taken a pivotal
role in influencing the salesperson's sales practices and customer's purchase
decisions.This study tries to examine the influence of retail operations and
hard-selling startegies on the relationship between salesperson's interpersonal
skills and customer repurchase intention. Salesperson's interpersonal skills
are the trained and tacit competencies that a salesperson employs to improve
customer relationship and sales performance. Many organizations prefer skill
training fails to attract repetitive purchases due to unavoidable extraneous
factors. It has become a necessity to understand the role of extraneous factors
like retail operations on the relationship between salesperson's interpersonal
skills and customer repurchase intention. The findings suggest that retail
operations significantly moderate the relationship between salesperson's
interpersonal skills and customer repurchase intention. We also find that
hard-sell sales techniques play a significant moderating role in negatively
infleuncing customer repurchase intention . This study has important
implications for retailers and sales managers.
| stat |
Graph-Based Continual Learning | Despite significant advances, continual learning models still suffer from
catastrophic forgetting when exposed to incrementally available data from
non-stationary distributions. Rehearsal approaches alleviate the problem by
maintaining and replaying a small episodic memory of previous samples, often
implemented as an array of independent memory slots. In this work, we propose
to augment such an array with a learnable random graph that captures pairwise
similarities between its samples, and use it not only to learn new tasks but
also to guard against forgetting. Empirical results on several benchmark
datasets show that our model consistently outperforms recently proposed
baselines for task-free continual learning.
| stat |
Nonparametric estimation of highest density regions for COVID-19 | Highest density regions refer to level sets containing points of relatively
high density. Their estimation from a random sample, generated from the
underlying density, allows to determine the clusters of the corresponding
distribution. This task can be accomplished considering different nonparametric
perspectives. From a practical point of view, reconstructing highest density
regions can be interpreted as a way of determining hot-spots, a crucial task
for understanding COVID-19 space-time evolution. In this work, we compare the
behavior of classical plug-in methods and a recently proposed hybrid algorithm
for highest density regions estimation through an extensive simulation study.
Both methodologies are applied to analyze a real data set about COVID-19 cases
in the United States.
| stat |
Learning Energy-based Model with Flow-based Backbone by Neural Transport
MCMC | Learning energy-based model (EBM) requires MCMC sampling of the learned model
as the inner loop of the learning algorithm. However, MCMC sampling of EBM in
data space is generally not mixing, because the energy function, which is
usually parametrized by deep network, is highly multi-modal in the data space.
This is a serious handicap for both the theory and practice of EBM. In this
paper, we propose to learn EBM with a flow-based model serving as a backbone,
so that the EBM is a correction or an exponential tilting of the flow-based
model. We show that the model has a particularly simple form in the space of
the latent variables of the flow-based model, and MCMC sampling of the EBM in
the latent space, which is a simple special case of neural transport MCMC,
mixes well and traverses modes in the data space. This enables proper sampling
and learning of EBM.
| stat |
Fractal Gaussian Networks: A sparse random graph model based on Gaussian
Multiplicative Chaos | We propose a novel stochastic network model, called Fractal Gaussian Network
(FGN), that embodies well-defined and analytically tractable fractal
structures. Such fractal structures have been empirically observed in diverse
applications. FGNs interpolate continuously between the popular purely random
geometric graphs (a.k.a. the Poisson Boolean network), and random graphs with
increasingly fractal behavior. In fact, they form a parametric family of sparse
random geometric graphs that are parametrized by a fractality parameter $\nu$
which governs the strength of the fractal structure. FGNs are driven by the
latent spatial geometry of Gaussian Multiplicative Chaos (GMC), a canonical
model of fractality in its own right. We asymptotically characterize the
expected number of edges and triangle in FGNs. We then examine the natural
question of detecting the presence of fractality and the problem of parameter
estimation based on observed network data, in addition to fundamental
properties of the FGN as a random graph model. We also explore fractality in
community structures by unveiling a natural stochastic block model in the
setting of FGNs.
| stat |
Bagging cross-validated bandwidth selection in nonparametric regression
estimation with applications to large-sized samples | Cross-validation is a well-known and widely used bandwidth selection method
in nonparametric regression estimation. However, this technique has two
remarkable drawbacks: (i) the large variability of the selected bandwidths, and
(ii) the inability to provide results in a reasonable time for very large
sample sizes. To overcome these problems, bagging cross-validation bandwidths
are analyzed in this paper. This approach consists in computing the
cross-validation bandwidths for a finite number of subsamples and then
rescaling the averaged smoothing parameters to the original sample size. Under
a random-design regression model, asymptotic expressions up to a second-order
for the bias and variance of the leave-one-out cross-validation bandwidth for
the Nadaraya--Watson estimator are obtained. Subsequently, the asymptotic bias
and variance and the limit distribution are derived for the bagged
cross-validation selector. Suitable choices of the number of subsamples and the
subsample size lead to an $n^{-1/2}$ rate for the convergence in distribution
of the bagging cross-validation selector, outperforming the rate $n^{-3/10}$ of
leave-one-out cross-validation. Several simulations and an illustration on a
real dataset related to the COVID-19 pandemic show the behavior of our proposal
and its better performance, in terms of statistical efficiency and computing
time, when compared to leave-one-out cross-validation.
| stat |
Testing for Network and Spatial Autocorrelation | Testing for dependence has been a well-established component of spatial
statistical analyses for decades. In particular, several popular test
statistics have desirable properties for testing for the presence of spatial
autocorrelation in continuous variables. In this paper we propose two
contributions to the literature on tests for autocorrelation. First, we propose
a new test for autocorrelation in categorical variables. While some methods
currently exist for assessing spatial autocorrelation in categorical variables,
the most popular method is unwieldy, somewhat ad hoc, and fails to provide
grounds for a single omnibus test. Second, we discuss the importance of testing
for autocorrelation in data sampled from the nodes of a network, motivated by
social network applications. We demonstrate that our proposed statistic for
categorical variables can both be used in the spatial and network setting.
| stat |
CauchyCP: a powerful test under non-proportional hazards using Cauchy
combination of change-point Cox regressions | Non-proportional hazards data are routinely encountered in randomized
clinical trials. In such cases, classic Cox proportional hazards model can
suffer from severe power loss, with difficulty in interpretation of the
estimated hazard ratio since the treatment effect varies over time. We propose
CauchyCP, an omnibus test of change-point Cox regression models, to overcome
both challenges while detecting signals of non-proportional hazards patterns.
Extensive simulation studies demonstrate that, compared to existing treatment
comparison tests under non-proportional hazards, the proposed CauchyCP test 1)
controls the type I error better at small $\alpha$ levels ($< 0.01$); 2)
increases the power of detecting time-varying effects; and 3) is more
computationally efficient. The superior performance of CauchyCP is further
illustrated using retrospective analyses of two randomized clinical trial
datasets and a pharmacogenetic biomarker study dataset. The R package
$\textit{CauchyCP}$ is publicly available on CRAN.
| stat |
Split regression modeling | In this note we study the benefits of splitting variables variables for
reducing the variance of linear functions of the regression coefficient
estimate. We show that splitting combined with shrinkage can result in
estimators with smaller mean squared error compared to popular shrinkage
estimators such as Lasso, ridge regression and garrote.
| stat |
On Deep Set Learning and the Choice of Aggregations | Recently, it has been shown that many functions on sets can be represented by
sum decompositions. These decompositons easily lend themselves to neural
approximations, extending the applicability of neural nets to set-valued
inputs---Deep Set learning. This work investigates a core component of Deep Set
architecture: aggregation functions. We suggest and examine alternatives to
commonly used aggregation functions, including learnable recurrent aggregation
functions. Empirically, we show that the Deep Set networks are highly sensitive
to the choice of aggregation functions: beyond improved performance, we find
that learnable aggregations lower hyper-parameter sensitivity and generalize
better to out-of-distribution input size.
| stat |
A Hybrid Approximation to the Marginal Likelihood | Computing the marginal likelihood or evidence is one of the core challenges
in Bayesian analysis. While there are many established methods for estimating
this quantity, they predominantly rely on using a large number of posterior
samples obtained from a Markov Chain Monte Carlo (MCMC) algorithm. As the
dimension of the parameter space increases, however, many of these methods
become prohibitively slow and potentially inaccurate. In this paper, we propose
a novel method in which we use the MCMC samples to learn a high probability
partition of the parameter space and then form a deterministic approximation
over each of these partition sets. This two-step procedure, which constitutes
both a probabilistic and a deterministic component, is termed a Hybrid
approximation to the marginal likelihood. We demonstrate its versatility in a
plethora of examples with varying dimension and sample size, and we also
highlight the Hybrid approximation's effectiveness in situations where there is
either a limited number or only approximate MCMC samples available.
| stat |
The causal effect of a timeout at stopping an opposing run in the NBA | In the summer of 2017, the National Basketball Association reduced the number
of total timeouts, along with other rule changes, to regulate the flow of the
game. With these rule changes, it becomes increasingly important for coaches to
effectively manage their timeouts. Understanding the utility of a timeout under
various game scenarios, e.g., during an opposing team's run, is of the utmost
importance. There are two schools of thought when the opposition is on a run:
(1) call a timeout and allow your team to rest and regroup, or (2) save a
timeout and hope your team can make corrections during play. This paper
investigates the credence of these tenets using the Rubin causal model
framework to quantify the causal effect of a timeout in the presence of an
opposing team's run. Too often overlooked, we carefully consider the stable
unit-treatment-value assumption (SUTVA) in this context and use SUTVA to
motivate our definition of units. To measure the effect of a timeout, we
introduce a novel, interpretable outcome based on the score difference to
describe broad changes in the scoring dynamics. This outcome is well-suited for
situations where the quantity of interest fluctuates frequently, a commonality
in many sports analytics applications. We conclude from our analysis that while
comebacks frequently occur after a run, it is slightly disadvantageous to call
a timeout during a run by the opposing team and further demonstrate that the
magnitude of this effect varies by franchise.
| stat |
Classifying histograms of medical data using information geometry of
beta distributions | In this paper, we use tools of information geometry to compare, average and
classify histograms. Beta distributions are fitted to the histograms and the
corresponding Fisher information geometry is used for comparison. We show that
this geometry is negatively curved, which guarantees uniqueness of the notion
of mean, and makes it suitable to classify histograms through the popular
K-means algorithm. We illustrate the use of these geometric tools in supervised
and unsupervised classification procedures of two medical data-sets, cardiac
shape deformations for the detection of pulmonary hypertension and brain
cortical thickness for the diagnosis of Alzheimer's disease.
| stat |
Robust and integrative Bayesian neural networks for likelihood-free
parameter inference | State-of-the-art neural network-based methods for learning summary statistics
have delivered promising results for simulation-based likelihood-free parameter
inference. Existing approaches require density estimation as a post-processing
step building upon deterministic neural networks, and do not take network
prediction uncertainty into account. This work proposes a robust integrated
approach that learns summary statistics using Bayesian neural networks, and
directly estimates the posterior density using categorical distributions. An
adaptive sampling scheme selects simulation locations to efficiently and
iteratively refine the predictive posterior of the network conditioned on
observations. This allows for more efficient and robust convergence on
comparatively large prior spaces. We demonstrate our approach on benchmark
examples and compare against related methods.
| stat |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.