title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Validating Gravity-Based Market Share Models Using Large-Scale
Transactional Data | Customer patronage behavior has been widely studied in market share modeling
contexts, which is an essential step towards modeling and solving competitive
facility location problems. Existing studies have conducted surveys to estimate
merchants' market share and factors of attractiveness to use in various
proposed mathematical models. Recent trends in Big Data analysis allow us to
better understand human behavior and decision making, potentially leading to
location models with more realistic assumptions. In this paper, we propose a
novel approach for validating Huff gravity market share model, using a
large-scale transactional dataset that describes customer patronage behavior in
a regional scale. Although the Huff model has been well-studied and widely used
in the context of competitive facility location and demand allocation, this
paper is the first in validating the Huff model with a real dataset. Our
approach helps to easily apply the model in different regions and with
different merchant categories. Experimental results show that the Huff model
fits well when modeling customer shopping behavior for a number of shopping
categories including grocery stores, clothing stores, gas stations, and
restaurants. We also conduct regression analysis to show that certain features
such as gender diversity and marital status diversity lead to stronger
validation of the Huff model. We believe we provide strong evidence, with the
help of real world data, that gravity-based market share models are viable
assumptions for competitive facility location models.
| stat |
Deep Probabilistic Graphical Modeling | Probabilistic graphical modeling (PGM) provides a framework for formulating
an interpretable generative process of data and expressing uncertainty about
unknowns, but it lacks flexibility. Deep learning (DL) is an alternative
framework for learning from data that has achieved great empirical success in
recent years. DL offers great flexibility, but it lacks the interpretability
and calibration of PGM. This thesis develops deep probabilistic graphical
modeling (DPGM.) DPGM consists in leveraging DL to make PGM more flexible. DPGM
brings about new methods for learning from data that exhibit the advantages of
both PGM and DL.
We use DL within PGM to build flexible models endowed with an interpretable
latent structure. One model class we develop extends exponential family PCA
using neural networks to improve predictive performance while enforcing the
interpretability of the latent factors. Another model class we introduce
enables accounting for long-term dependencies when modeling sequential data,
which is a challenge when using purely DL or PGM approaches. Finally, DPGM
successfully solves several outstanding problems of probabilistic topic models,
a widely used family of models in PGM.
DPGM also brings about new algorithms for learning with complex data. We
develop reweighted expectation maximization, an algorithm that unifies several
existing maximum likelihood-based algorithms for learning models parameterized
by neural networks. This unifying view is made possible using expectation
maximization, a canonical inference algorithm in PGM. We also develop
entropy-regularized adversarial learning, a learning paradigm that deviates
from the traditional maximum likelihood approach used in PGM. From the DL
perspective, entropy-regularized adversarial learning provides a solution to
the long-standing mode collapse problem of generative adversarial networks, a
widely used DL approach.
| stat |
Gaussian Process for Functional Data Analysis: The GPFDA Package for R | We present and describe the GPFDA package for R. The package provides
flexible functionalities for dealing with Gaussian process regression (GPR)
models for functional data. Multivariate functional data, functional data with
multidimensional inputs, and nonseparable and/or nonstationary covariance
structures can be modeled. In addition, the package fits functional regression
models where the mean function depends on scalar and/or functional covariates
and the covariance structure is modeled by a GPR model. In this paper, we
present the versatility of GPFDA with respect to mean function and covariance
function specifications and illustrate the implementation of estimation and
prediction of some models through reproducible numerical examples.
| stat |
A hypothesis test of feasibility for external pilot trials assessing
recruitment, follow-up and adherence rates | The power of a large clinical trial can be adversely affected by low
recruitment, follow-up and adherence rates. External pilot trials estimate
these rates and use them, via pre-specified decision rules, to determine if the
definitive trial is feasible and should go ahead. There is little
methodological research underpinning how these decision rules, or the sample
size of the pilot, should be chosen. In this paper we propose a hypothesis test
of the feasibility of a definitive trial, to be applied to the external pilot
data and used to make progression decisions. We quantify feasibility by the
power of the planned trial, as a function of recruitment, follow-up and
adherence rates. We use this measure to define hypotheses to test in the pilot,
propose a test statistic, and show how the error rates of this test can be
calculated for the common scenario of a two-arm parallel group definitive trial
with a single normally distributed primary endpoint. We use our method to
re-design TIGA-CUB, an external pilot trial comparing a psychotherapy with
treatment as usual for children with conduct disorders. We then extend our
formulation to include using the pilot data to estimate the standard deviation
of the primary endpoint. and incorporate this into the progression decision.
| stat |
A Le Cam Type Bound for Adversarial Learning and Applications | Robustness of machine learning methods is essential for modern practical
applications. Given the arms race between attack and defense methods, one may
be curious regarding the fundamental limits of any defense mechanism. In this
work, we focus on the problem of learning from noise-injected data, where the
existing literature falls short by either assuming a specific attack method or
by over-specifying the learning problem. We shed light on the
information-theoretic limits of adversarial learning without assuming a
particular learning process or attacker. Finally, we apply our general bounds
to a canonical set of non-trivial learning problems and provide examples of
common types of attacks.
| stat |
Efficient nonparametric inference on the effects of stochastic
interventions under two-phase sampling, with applications to vaccine efficacy
trials | The advent and subsequent widespread availability of preventive vaccines has
altered the course of public health over the past century. Despite this
success, effective vaccines to prevent many high-burden diseases, including
HIV, have been slow to develop. Vaccine development can be aided by the
identification of immune response markers that serve as effective surrogates
for clinically significant infection or disease endpoints. However, measuring
immune response marker activity is often costly, which has motivated the usage
of two-phase sampling for immune response evaluation in clinical trials of
preventive vaccines. In such trials, the measurement of immunological markers
is performed on a subset of trial participants, where enrollment in this second
phase is potentially contingent on the observed study outcome and other
participant-level information. We propose nonparametric methodology for
efficiently estimating a counterfactual parameter that quantifies the impact of
a given immune response marker on the subsequent probability of infection.
Along the way, we fill in theoretical gaps pertaining to the asymptotic
behavior of nonparametric efficient estimators in the context of two-phase
sampling, including a multiple robustness property enjoyed by our estimators.
Techniques for constructing confidence intervals and hypothesis tests are
presented, and an open source software implementation of the methodology, the
txshift R package, is introduced. We illustrate the proposed techniques using
data from a recent preventive HIV vaccine efficacy trial.
| stat |
On deletion diagnostic statistic in regression | A brief review about choosing a normalizing or scaling matrix for deletion
diagnostic statistic in regression is made. Some results and comments are
added.
| stat |
Harmonic Decompositions of Convolutional Networks | We present a description of the function space and the smoothness class
associated with a convolutional network using the machinery of reproducing
kernel Hilbert spaces. We show that the mapping associated with a convolutional
network expands into a sum involving elementary functions akin to spherical
harmonics. This functional decomposition can be related to the functional ANOVA
decomposition in nonparametric statistics. Building off our functional
characterization of convolutional networks, we obtain statistical bounds
highlighting an interesting trade-off between the approximation error and the
estimation error.
| stat |
$hv$-Block Cross Validation is not a BIBD: a Note on the Paper by Jeff
Racine (2000) | This note corrects a mistake in the paper "consistent cross-validatory
model-selection for dependent data: $hv$-block cross-validation" by Racine
(2000). In his paper, he implied that the therein proposed $hv$-block
cross-validation is consistent in the sense of Shao (1993). To get this
intuition, he relied on the speculation that $hv$-block is a balanced
incomplete block design (BIBD). This note demonstrates that this is not the
case, and thus the theoretical consistency of $hv$-block remains an open
question. In addition, I also provide a Python program counting the number of
occurrences of each sample and each pair of samples.
| stat |
On the optimality of kernels for high-dimensional clustering | This paper studies the optimality of kernel methods in high-dimensional data
clustering. Recent works have studied the large sample performance of kernel
clustering in the high-dimensional regime, where Euclidean distance becomes
less informative. However, it is unknown whether popular methods, such as
kernel k-means, are optimal in this regime. We consider the problem of
high-dimensional Gaussian clustering and show that, with the exponential kernel
function, the sufficient conditions for partial recovery of clusters using the
NP-hard kernel k-means objective matches the known information-theoretic limit
up to a factor of $\sqrt{2}$ for large $k$. It also exactly matches the known
upper bounds for the non-kernel setting. We also show that a semi-definite
relaxation of the kernel k-means procedure matches up to constant factors, the
spectral threshold, below which no polynomial-time algorithm is known to
succeed. This is the first work that provides such optimality guarantees for
the kernel k-means as well as its convex relaxation. Our proofs demonstrate the
utility of the less known polynomial concentration results for random variables
with exponentially decaying tails in a higher-order analysis of kernel methods.
| stat |
Exact Confidence Intervals for Linear Combinations of Multinomial
Probabilities | Linear combinations of multinomial probabilities, such as those resulting
from contingency tables, are of use when evaluating classification system
performance. While large sample inference methods for these combinations exist,
small sample methods exist only for regions on the multinomial parameter space
instead of the linear combinations. However, in medical classification problems
it is common to have small samples necessitating a small sample confidence
interval on linear combinations of multinomial probabilities. Therefore, in
this paper we derive an exact confidence interval, through the use of fiducial
inference, for linear combinations of multinomial probabilities. Simulation
demonstrates the presented interval's adherence to exact coverage.
Additionally, an adjustment to the exact interval is provided, giving shorter
lengths while still achieving better coverage than large sample methods.
Computational efficiencies in estimation of the exact interval are achieved
through the application of a fast Fourier transform and combining a numerical
solver and stochastic optimizer to find solutions. The exact confidence
interval presented in this paper allows for comparisons between diagnostic
methods previously unavailable, demonstrated through an example of diagnosing
chronic allograph nephropathy in post kidney transplant patients.
| stat |
Pseudo-Rehearsal for Continual Learning with Normalizing Flows | Catastrophic forgetting (CF) happens whenever a neural network overwrites
past knowledge while being trained on new tasks. Common techniques to handle CF
include regularization of the weights (using, e.g., their importance on past
tasks), and rehearsal strategies, where the network is constantly re-trained on
past data. Generative models have also been applied for the latter, in order to
have endless sources of data. In this paper, we propose a novel method that
combines the strengths of regularization and generative-based rehearsal
approaches. Our generative model consists of a normalizing flow (NF), a
probabilistic and invertible neural network, trained on the internal embeddings
of the network. By keeping a single NF conditioned on the task, we show that
our memory overhead remains constant. In addition, exploiting the invertibility
of the NF, we propose a simple approach to regularize the network's embeddings
with respect to past tasks. We show that our method performs favorably with
respect to state-of-the-art approaches in the literature, with bounded
computational power and memory overheads.
| stat |
Integration of Survival Data from Multiple Studies | We introduce a statistical procedure that integrates survival data from
multiple biomedical studies, to improve the accuracy of predictions of survival
or other events, based on individual clinical and genomic profiles, compared to
models developed leveraging only a single study or meta-analytic methods. The
method accounts for potential differences in the relation between predictors
and outcomes across studies, due to distinct patient populations, treatments
and technologies to measure outcomes and biomarkers. These differences are
modeled explicitly with study-specific parameters. We use hierarchical
regularization to shrink the study-specific parameters towards each other and
to borrow information across studies. Shrinkage of the study-specific
parameters is controlled by a similarity matrix, which summarizes differences
and similarities of the relations between covariates and outcomes across
studies. We illustrate the method in a simulation study and using a collection
of gene-expression datasets in ovarian cancer. We show that the proposed model
increases the accuracy of survival prediction compared to alternative
meta-analytic methods.
| stat |
Dynamic Principal Subspaces with Sparsity in High Dimensions | Principal component analysis (PCA) is a versatile tool to reduce the
dimensionality which has wide applications in statistics and machine learning
community. It is particularly useful to model data in high-dimensional
scenarios where the number of variables $p$ is comparable to, or much larger
than the sample size $n$. Despite extensive literature on this topic,
researches have focused on modeling static principal eigenvectors or subspaces,
which is unsuitable for stochastic processes that are dynamic in nature. To
characterize the change in the whole course of high-dimensional data
collection, we propose a unified framework to estimate dynamic principal
subspaces spanned by leading eigenvectors of covariance matrices. In the
proposed framework, we formulate an optimization problem by combining the
kernel smoothing and regularization penalty together with the orthogonality
constraint, which can be effectively solved by the proximal gradient method for
manifold optimization. We show that our method is suitable for high-dimensional
data observed under both common and irregular designs. In addition, theoretical
properties of the estimators are investigated under $l_q (0 \leq q \leq 1)$
sparsity. Extensive experiments demonstrate the effectiveness of the proposed
method in both simulated and real data examples.
| stat |
Non-Smooth Backfitting for Excess Risk Additive Regression Model with
Two Survival Time-Scales | We present a new backfitting algorithm estimating the complex structured
non-parametric survival model of Scheike (2001) without having to use
smoothing. The considered model is a non-parametric survival model with two
time-scales that are equivalent up to a constant that varies over the subjects.
Covariate effects are modelled linearly on each time scale by additive Aalen
models. Estimators of the cumulative intensities on the two time-scales are
suggested by solving local estimating equations jointly on the two time-scales.
We are able to estimate the cumulative intensities solving backfitting
estimating equations without using smoothing methods and we provide large
sample properties and simultaneous confidence bands. The model is applied to
data on myocardial infarction providing a separation of the two effects
stemming from time since diagnosis and age.
| stat |
Discriminating between and within (semi)continuous classes of both
Tweedie and geometric Tweedie models | In both Tweedie and geometric Tweedie models, the common power parameter
$p\notin(0,1)$ works as an automatic distribution selection. It mainly
separates two subclasses of semicontinuous ($1<p<2$) and positive continuous
($p\geq 2$) distributions. Our paper centers around exploring diagnostic tools
based on the maximum likelihood ratio test and minimum Kolmogorov-Smirnov
distance methods in order to discriminate very close distributions within each
subclass of these two models according to values of $p$. Grounded on the unique
equality of variation indices, we also discriminate the gamma and geometric
gamma distributions with $p=2$ in Tweedie and geometric Tweedie families,
respectively. Probabilities of correct selection for several combinations of
dispersion parameters, means and sample sizes are examined by simulations. We
thus perform a numerical comparison study to assess the discrimination
procedures in these subclasses of two families. Finally, semicontinuous
($1<p\leq 2$) distributions in the broad sense are significantly more
distinguishable than the over-varied continuous ($p>2$) ones; and two datasets
for illustration purposes are investigated.
| stat |
Generalizing Variational Autoencoders with Hierarchical Empirical Bayes | Variational Autoencoders (VAEs) have experienced recent success as
data-generating models by using simple architectures that do not require
significant fine-tuning of hyperparameters. However, VAEs are known to suffer
from over-regularization which can lead to failure to escape local maxima. This
phenomenon, known as posterior collapse, prevents learning a meaningful latent
encoding of the data. Recent methods have mitigated this issue by
deterministically moment-matching an aggregated posterior distribution to an
aggregate prior. However, abandoning a probabilistic framework (and thus
relying on point estimates) can both lead to a discontinuous latent space and
generate unrealistic samples. Here we present Hierarchical Empirical Bayes
Autoencoder (HEBAE), a computationally stable framework for probabilistic
generative models. Our key contributions are two-fold. First, we make gains by
placing a hierarchical prior over the encoding distribution, enabling us to
adaptively balance the trade-off between minimizing the reconstruction loss
function and avoiding over-regularization. Second, we show that assuming a
general dependency structure between variables in the latent space produces
better convergence onto the mean-field assumption for improved posterior
inference. Overall, HEBAE is more robust to a wide-range of hyperparameter
initializations than an analogous VAE. Using data from MNIST and CelebA, we
illustrate the ability of HEBAE to generate higher quality samples based on FID
score than existing autoencoder-based approaches.
| stat |
Regularising Generalised Linear Mixed Models with an autoregressive
random effect | We address regularised versions of the Expectation-Maximisation (EM)
algorithm for Generalised Linear Mixed Models (GLMM) in the context of panel
data (measured on several individuals at different time-points). A random
response y is modelled by a GLMM, using a set X of explanatory variables and
two random effects. The first one introduces the dependence within individuals
on which data is repeatedly collected while the second one embodies the
serially correlated time-specific effect shared by all the individuals.
Variables in X are assumed many and redundant, so that regression demands
regularisation. In this context, we first propose a L2-penalised EM algorithm,
and then a supervised component-based regularised EM algorithm as an
alternative.
| stat |
Generalizing trial evidence to target populations in non-nested designs:
Applications to AIDS clinical trials | Comparative effectiveness evidence from randomized trials may not be directly
generalizable to a target population of substantive interest when, as in most
cases, trial participants are not randomly sampled from the target population.
Motivated by the need to generalize evidence from two trials conducted in the
AIDS Clinical Trials Group (ACTG), we consider weighting, regression and doubly
robust estimators to estimate the causal effects of HIV interventions in a
specified population of people living with HIV in the USA. We focus on a
non-nested trial design and discuss strategies for both point and variance
estimation of the target population average treatment effect. Specifically in
the generalizability context, we demonstrate both analytically and empirically
that estimating the known propensity score in trials does not increase the
variance for each of the weighting, regression and doubly robust estimators. We
apply these methods to generalize the average treatment effects from two ACTG
trials to specified target populations and operationalize key practical
considerations. Finally, we report on a simulation study that investigates the
finite-sample operating characteristics of the generalizability estimators and
their sandwich variance estimators.
| stat |
Mean shift cluster recognition method implementation in the nested
sampling algorithm | Nested sampling is an efficient algorithm for the calculation of the Bayesian
evidence and posterior parameter probability distributions. It is based on the
step-by-step exploration of the parameter space by Monte Carlo sampling with a
series of values sets called live points that evolve towards the region of
interest, i.e. where the likelihood function is maximal. In presence of several
local likelihood maxima, the algorithm converges with difficulty. Some
systematic errors can also be introduced by unexplored parameter volume
regions. In order to avoid this, different methods are proposed in the
literature for an efficient search of new live points, even in presence of
local maxima. Here we present a new solution based on the mean shift cluster
recognition method implemented in a random walk search algorithm. The
clustering recognition is integrated within the Bayesian analysis program
NestedFit. It is tested with the analysis of some difficult cases. Compared to
the analysis results without cluster recognition, the computation time is
considerably reduced. At the same time, the entire parameter space is
efficiently explored, which translates into a smaller uncertainty of the
extracted value of the Bayesian evidence.
| stat |
Relaxed-Responsibility Hierarchical Discrete VAEs | Successfully training Variational Autoencoders (VAEs) with a hierarchy of
discrete latent variables remains an area of active research.
Vector-Quantised VAEs are a powerful approach to discrete VAEs, but naive
hierarchical extensions can be unstable when training. Leveraging insights from
classical methods of inference we introduce \textit{Relaxed-Responsibility
Vector-Quantisation}, a novel way to parameterise discrete latent variables, a
refinement of relaxed Vector-Quantisation that gives better performance and
more stable training. This enables a novel approach to hierarchical discrete
variational autoencoders with numerous layers of latent variables (here up to
32) that we train end-to-end. Within hierarchical probabilistic deep generative
models with discrete latent variables trained end-to-end, we achieve
state-of-the-art bits-per-dim results for various standard datasets. % Unlike
discrete VAEs with a single layer of latent variables, we can produce samples
by ancestral sampling: it is not essential to train a second autoregressive
generative model over the learnt latent representations to then sample from and
then decode. % Moreover, that latter approach in these deep hierarchical models
would require thousands of forward passes to generate a single sample. Further,
we observe different layers of our model become associated with different
aspects of the data.
| stat |
Oracle lower bounds for stochastic gradient sampling algorithms | We consider the problem of sampling from a strongly log-concave density in
$\mathbb{R}^d$, and prove an information theoretic lower bound on the number of
stochastic gradient queries of the log density needed. Several popular sampling
algorithms (including many Markov chain Monte Carlo methods) operate by using
stochastic gradients of the log density to generate a sample; our results
establish an information theoretic limit for all these algorithms.
We show that for every algorithm, there exists a well-conditioned strongly
log-concave target density for which the distribution of points generated by
the algorithm would be at least $\epsilon$ away from the target in total
variation distance if the number of gradient queries is less than
$\Omega(\sigma^2 d/\epsilon^2)$, where $\sigma^2 d$ is the variance of the
stochastic gradient. Our lower bound follows by combining the ideas of Le Cam
deficiency routinely used in the comparison of statistical experiments along
with standard information theoretic tools used in lower bounding Bayes risk
functions. To the best of our knowledge our results provide the first
nontrivial dimension-dependent lower bound for this problem.
| stat |
Contextual Stress and Maternal Sensitivity: A Meta-Analytic Review of
Stress Associations with the Maternal Behavior Q-Sort in Observational
Studies | Maternal sensitivity is a modifiable determinant of infant attachment
security and a precursor to optimal child development. Contextual stressors
undermine sensitivity, but research was yet to be synthesized. We aimed to
identify i) types of stress associations analyzed in studies of maternal
sensitivity and ii) the strength of effects of various stress factors. A
systematic search identified all studies that used the Maternal Behavior Q-Sort
(MBQS) to code sensitivity in dyadic observations and that reported a
coefficient for MBQS associations with contextual stress. Identified stressors
cohered around three spheres: sociodemography (maternal education, family
income, composite SES, maternal age and cohabitation status); parenting stress
(perceived maternal stress related to parenting); and mental health
(specifically maternal internalizing symptoms). Seven meta-analyses (combined
ns range 223-1239) of a subset of 30 effects from 20 articles, and a
multi-level meta-analysis (N=1324) assessed aggregated correlations with
sensitivity. Significant mean effects emerged in expected directions, whereby
all stress indicators were negatively associated with sensitivity. Small
effects were found for associations with parenting stress (r=-0.13) and mental
health indicators (r=-0.12). Generally moderate effects were found for
associations with socio-demographic indicators (range r=-0.12 to r=0.32).
Emerging findings support the proposition that in various contexts of stress,
maternal sensitivity to infant needs can be undermined. Implications and
research directions are discussed.
| stat |
Introducing a Generative Adversarial Network Model for Lagrangian
Trajectory Simulation | We introduce a generative adversarial network (GAN) model to simulate the
3-dimensional Lagrangian motion of particles trapped in the recirculation zone
of a buoyancy-opposed flame. The GAN model comprises a stochastic recurrent
neural network, serving as a generator, and a convoluted neural network,
serving as a discriminator. Adversarial training was performed to the point
where the best-trained discriminator failed to distinguish the ground truth
from the trajectory produced by the best-trained generator. The model
performance was then benchmarked against a statistical analysis performed on
both the simulated trajectories and the ground truth, with regard to the
accuracy and generalization criteria.
| stat |
Incorporating structured assumptions with probabilistic graphical models
in fMRI data analysis | With the wide adoption of functional magnetic resonance imaging (fMRI) by
cognitive neuroscience researchers, large volumes of brain imaging data have
been accumulated in recent years. Aggregating these data to derive scientific
insights often faces the challenge that fMRI data are high-dimensional,
heterogeneous across people, and noisy. These challenges demand the development
of computational tools that are tailored both for the neuroscience questions
and for the properties of the data. We review a few recently developed
algorithms in various domains of fMRI research: fMRI in naturalistic tasks,
analyzing full-brain functional connectivity, pattern classification, inferring
representational similarity and modeling structured residuals. These algorithms
all tackle the challenges in fMRI similarly: they start by making clear
statements of assumptions about neural data and existing domain knowledge,
incorporating those assumptions and domain knowledge into probabilistic
graphical models, and using those models to estimate properties of interest or
latent structures in the data. Such approaches can avoid erroneous findings,
reduce the impact of noise, better utilize known properties of the data, and
better aggregate data across groups of subjects. With these successful cases,
we advocate wider adoption of explicit model construction in cognitive
neuroscience. Although we focus on fMRI, the principle illustrated here is
generally applicable to brain data of other modalities.
| stat |
Variable Selection in Regression-based Estimation of Dynamic Treatment
Regimes | Dynamic treatment regimes (DTRs) consist of a sequence of decision rules, one
per stage of intervention, that finds effective treatments for individual
patients according to patient information history. DTRs can be estimated from
models which include the interaction between treatment and a small number of
covariates which are often chosen a priori. However, with increasingly large
and complex data being collected, it is difficult to know which prognostic
factors might be relevant in the treatment rule. Therefore, a more data-driven
approach of selecting these covariates might improve the estimated decision
rules and simplify models to make them easier to interpret. We propose a
variable selection method for DTR estimation using penalized dynamic weighted
least squares. Our method has the strong heredity property, that is, an
interaction term can be included in the model only if the corresponding main
terms have also been selected. Through simulations, we show our method has both
the double robustness property and the oracle property, and the newly proposed
methods compare favorably with other variable selection approaches.
| stat |
A new Granger causality measure for eliminating the confounding
influence of latent common inputs | In this paper, we propose a new Granger causality measure which is robust
against the confounding influence of latent common inputs. This measure is
inspired by partial Granger causality in the literature, and its variant. Using
numerical experiments we first show that the test statistics for detecting
directed interactions between time series approximately obey the
$F$-distributions when there are no interactions. Then, we propose a practical
procedure for inferring directed interactions, which is based on the idea of
multiple statistical test in situations where the confounding influence of
latent common inputs may exist. The results of numerical experiments
demonstrate that the proposed method successfully eliminates the influence of
latent common inputs while the normal Granger causality method detects spurious
interactions due to the influence of the confounder.
| stat |
Predicting Infection of COVID-19 in Japan: State Space Modeling Approach | The number of confirmed cases of the coronavirus disease (COVID-19) in Japan
has been increasing day by day and has had a serious impact on the society
especially after the declaration of the state of emergency on April 7, 2020.
This study analyzes the real time data from March 1 to April 22, 2020 by
adopting a sophisticated statistical modeling tool based on the state space
model combined with the well-known susceptible-exposed-infected (SIR) model.
The model estimation and forecasting are conducted using the Bayesian
methodology. The present study provides the parameter estimates of the unknown
parameters that critically determine the epidemic process derived from the SIR
model and prediction of the future transition of the infectious proportion
including the size and timing of the epidemic peak with the prediction
intervals that naturally accounts for the uncertainty. The prediction results
under various scenarios reveals that the temporary reduction in the infection
rate until the planned lifting of the state on May 6 will only delay the
epidemic peak slightly. In order to minimize the spread of the epidemic, it is
strongly suggested that an intervention is carried out for an extended period
of time and that the government and individuals make a long term effort to
reduce the infection rate even after the lifting.
| stat |
Predicting the Output From a Stochastic Computer Model When a
Deterministic Approximation is Available | The analysis of computer models can be aided by the construction of surrogate
models, or emulators, that statistically model the numerical computer model.
Increasingly, computer models are becoming stochastic, yielding different
outputs each time they are run, even if the same input values are used.
Stochastic computer models are more difficult to analyse and more difficult to
emulate - often requiring substantially more computer model runs to fit. We
present a method of using deterministic approximations of the computer model to
better construct an emulator. The method is applied to numerous toy examples,
as well as an idealistic epidemiology model, and a model from the building
performance field.
| stat |
Advantage of Deep Neural Networks for Estimating Functions with
Singularity on Curves | We develop a theory to elucidate the reason that deep neural networks (DNNs)
perform better than other methods. In terms of the nonparametric regression
problem, it is well known that many standard methods attain the minimax optimal
rate of estimation errors for smooth functions, and thus, it is not
straightforward to identify the theoretical advantages of DNNs. This study
fills this gap by considering the estimation for a class of non-smooth
functions with singularities on smooth curves. Our findings are as follows: (i)
We derive the generalization error of a DNN estimator and prove that its
convergence rate is almost optimal. (ii) We reveal that a certain class of
common models are sub-optimal, including linear estimators and other harmonic
analysis methods such as wavelets and curvelets. This advantage of DNNs comes
from a fact that a shape of singularity can be successfully handled by their
multi-layered structure.
| stat |
Optimizing Ensemble Weights and Hyperparameters of Machine Learning
Models for Regression Problems | Aggregating multiple learners through an ensemble of models aim to make
better predictions by capturing the underlying distribution of the data more
accurately. Different ensembling methods, such as bagging, boosting, and
stacking/blending, have been studied and adopted extensively in research and
practice. While bagging and boosting focus more on reducing variance and bias,
respectively, stacking approaches target both by finding the optimal way to
combine base learners. In stacking with the weighted average, ensembles are
created from weighted averages of multiple base learners. It is known that
tuning hyperparameters of each base learner inside the ensemble weight
optimization process can produce better performing ensembles. To this end, an
optimization-based nested algorithm that considers tuning hyperparameters as
well as finding the optimal weights to combine ensembles (Generalized Weighted
Ensemble with Internally Tuned Hyperparameters (GEM-ITH)) is designed. Besides,
Bayesian search was used to speed-up the optimizing process, and a heuristic
was implemented to generate diverse and well-performing base learners. The
algorithm is shown to be generalizable to real data sets through analyses with
ten publicly available data sets.
| stat |
Flexible Predictive Distributions from Varying-Thresholds Modelling | A general class of models is proposed that is able to estimate the whole
predictive distribution of a dependent variable $Y$ given a vector of
explanatory variables $\xb$. The models exploit that the strength of
explanatory variables to distinguish between low and high values of the
dependent variable may vary across the thresholds that are used to define low
and high. Simple linear versions of the models are generalizations of classical
linear regression models but also of widely used ordinal regression models.
They allow to visualize the effect of explanatory variables in the form of
parameter functions. More general models are based on efficient nonparametric
approaches like random forests, which are more flexible and are strong
prediction tools. A general estimation method is given that can use all the
estimation tools that have been proposed for binary regression, including
selection methods like the lasso or elastic net. For linearly structured models
maximum likelihood estimates are derived. The usefulness of the models is
illustrated by simulations and several real data set.
| stat |
Testing for publication bias in meta-analysis under Copas selection
model | In meta-analyses, publication bias is a well-known, important and challenging
issue because the validity of the results from a meta-analysis is threatened if
the sample of studies retrieved for review is biased. One popular method to
deal with publication bias is the Copas selection model, which provides a
flexible sensitivity analysis for correcting the estimates with considerable
insight into the data suppression mechanism. However, rigorous testing
procedures under the Copas selection model to detect bias are lacking. To fill
this gap, we develop a score-based test for detecting publication bias under
the Copas selection model. We reveal that the behavior of the standard score
test statistic is irregular because the parameters of the Copas selection model
disappear under the null hypothesis, leading to an identifiability problem. We
propose a novel test statistic and derive its limiting distribution. A
bootstrap procedure is provided to obtain the p-value of the test for practical
applications. We conduct extensive Monte Carlo simulations to evaluate the
performance of the proposed test and apply the method to several existing
meta-analyses.
| stat |
The Geometry of Over-parameterized Regression and Adversarial
Perturbations | Classical regression has a simple geometric description in terms of a
projection of the training labels onto the column space of the design matrix.
However, for over-parameterized models -- where the number of fit parameters is
large enough to perfectly fit the training data -- this picture becomes
uninformative. Here, we present an alternative geometric interpretation of
regression that applies to both under- and over-parameterized models. Unlike
the classical picture which takes place in the space of training labels, our
new picture resides in the space of input features. This new feature-based
perspective provides a natural geometric interpretation of the double-descent
phenomenon in the context of bias and variance, explaining why it can occur
even in the absence of label noise. Furthermore, we show that adversarial
perturbations -- small perturbations to the input features that result in large
changes in label values -- are a generic feature of biased models, arising from
the underlying geometry. We demonstrate these ideas by analyzing three minimal
models for over-parameterized linear least squares regression: without basis
functions (input features equal model features) and with linear or nonlinear
basis functions (two-layer neural networks with linear or nonlinear activation
functions, respectively).
| stat |
Structured Shrinkage Priors | In many regression settings the unknown coefficients may have some known
structure, for instance they may be ordered in space or correspond to a
vectorized matrix or tensor. At the same time, the unknown coefficients may be
sparse, with many nearly or exactly equal to zero. However, many commonly used
priors and corresponding penalties for coefficients do not encourage
simultaneously structured and sparse estimates. In this paper we develop
structured shrinkage priors that generalize multivariate normal, Laplace,
exponential power and normal-gamma priors. These priors allow the regression
coefficients to be correlated a priori without sacrificing elementwise sparsity
or shrinkage. The primary challenges in working with these structured shrinkage
priors are computational, as the corresponding penalties are intractable
integrals and the full conditional distributions that are needed to approximate
the posterior mode or simulate from the posterior distribution may be
non-standard. We overcome these issues using a flexible elliptical slice
sampling procedure, and demonstrate that these priors can be used to introduce
structure while preserving sparsity of the corresponding penalized estimate
given by the posterior mode.
| stat |
Spatially Adaptive Colocalization Analysis in Dual-Color Fluorescence
Microscopy | Colocalization analysis aims to study complex spatial associations between
bio-molecules via optical imaging techniques. However, existing colocalization
analysis workflows only assess an average degree of colocalization within a
certain region of interest and ignore the unique and valuable spatial
information offered by microscopy. In the current work, we introduce a new
framework for colocalization analysis that allows us to quantify colocalization
levels at each individual location and automatically identify pixels or regions
where colocalization occurs. The framework, referred to as spatially adaptive
colocalization analysis (SACA), integrates a pixel-wise local kernel model for
colocalization quantification and a multi-scale adaptive propagation-separation
strategy for utilizing spatial information to detect colocalization in a
spatially adaptive fashion. Applications to simulated and real biological
datasets demonstrate the practical merits of SACA in what we hope to be an
easily applicable and robust colocalization analysis method. In addition,
theoretical properties of SACA are investigated to provide rigorous statistical
justification.
| stat |
Sample size calculations for single-arm survival studies using
transformations of the Kaplan-Meier estimator | In single-arm clinical trials with survival outcomes, the Kaplan-Meier
estimator and its confidence interval are widely used to assess survival
probability and median survival time. Since the asymptotic normality of the
Kaplan-Meier estimator is a common result, the sample size calculation methods
have not been studied in depth. An existing sample size calculation method is
founded on the asymptotic normality of the Kaplan-Meier estimator using the log
transformation. However, the small sample properties of the log transformed
estimator are quite poor in small sample sizes (which are typical situations in
single-arm trials), and the existing method uses an inappropriate standard
normal approximation to calculate sample sizes. These issues can seriously
influence the accuracy of results. In this paper, we propose alternative
methods to determine sample sizes based on a valid standard normal
approximation with several transformations that may give an accurate normal
approximation even with small sample sizes. In numerical evaluations via
simulations, some of the proposed methods provided more accurate results, and
the empirical power of the proposed method with the arcsine square-root
transformation tended to be closer to a prescribed power than the other
transformations. These results were supported when methods were applied to data
from three clinical trials.
| stat |
Stochastic parameterization with VARX processes | In this study we investigate a data-driven stochastic methodology to
parameterize small-scale features in a prototype multiscale dynamical system,
the Lorenz '96 (L96) model. We propose to model the small-scale features using
a vector autoregressive process with exogenous variable (VARX), estimated from
given sample data. To reduce the number of parameters of the VARX we impose a
diagonal structure on its coefficient matrices. We apply the VARX to two
different configurations of the 2-layer L96 model, one with common parameter
choices giving unimodal invariant probability distributions for the L96 model
variables, and one with non-standard parameters giving trimodal distributions.
We show through various statistical criteria that the proposed VARX performs
very well for the unimodal configuration, while keeping the number of
parameters linear in the number of model variables. We also show that the
parameterization performs accurately for the very challenging trimodal L96
configuration by allowing for a dense (non-diagonal) VARX covariance matrix.
| stat |
An Instrumental Variable Estimator for Mixed Indicators: Analytic
Derivatives and Alternative Parameterizations | Methodological development of the Model-implied Instrumental Variable (MIIV)
estimation framework has proved fruitful over the last three decades. Major
milestones include Bollen's (1996) original development of the MIIV estimator
and its robustness properties for continuous endogenous variable SEMs, the
extension of the MIIV estimator to ordered categorical endogenous variables
(Bollen \& Maydeu-Olivares, 2007), and the introduction of a Generalized Method
of Moments (GMM) estimator (Bollen, Kolenikov \& Bauldry, 2014). This paper
furthers these developments by making several unique contributions not present
in the prior literature: (1) we use matrix calculus to derive the analytic
derivatives of the PIV estimator, (2) we extend the PIV estimator to apply to
any mixture of binary, ordinal, and continuous variables, (3) we generalize the
PIV model to include intercepts and means, (4) we devise a method to input
known threshold values for ordinal observed variables, and (5) we enable a
general parameterization that permits the estimation of means, variances, and
covariances of the underlying variables to use as input into a SEM analysis
with PIV. An empirical example illustrates a mixture of continuous variables
and ordinal variables with fixed thresholds. We also include a simulation study
to compare the performance of this novel estimator to WLSMV.
| stat |
Achieving higher taxi outflows from a congested drop-off lane: a
simulation-based policy study | We examine special lanes used by taxis and other shared-ride services to
drop-off patrons at airport and rail terminals. Vehicles are prohibited from
overtaking each other within the lane. They must therefore wait in a
first-in-first-out queue during busy periods. Patrons are often discharged from
vehicles only upon reaching a desired drop-off area near the terminal entrance.
When wait times grow long, however, some vehicles discharge their patrons in
advance of that desired area. A train station in Eastern China is selected as a
case study. Its FIFO drop-off lane is presently managed by policemen who allow
taxis to enter the lane in batched fashion. Inefficiencies are observed because
curb space near the terminal often goes unused. This is true even when
supplemental batches of taxis are released into the lane in efforts to fill
those spaces. A microscopic simulation model of a FIFO drop-off lane is
developed in-house, and is painstakingly calibrated to data measured at the
study site. Simulation experiments indicate that rescinding the FIFO lane's
present batching strategy can increase taxi outflow by more than 26 percent.
Further experiments show that even greater gains can be achieved by batching
taxis, but requiring them to discharge patrons when forced by downstream queues
to stop a prescribed distance in advance of a desired drop-off area. Further
gains were predicted by requiring the lead taxi in each batch to discharge its
patron only after travelling a prescribed distance beyond a desired location.
Practical implications are discussed in light of the present boom in
shared-ride services.
| stat |
Modeling spatial data using local likelihood estimation and a Mat\'ern
to SAR translation | Modeling data with non-stationary covariance structure is important to
represent heterogeneity in geophysical and other environmental spatial
processes. In this work, we investigate a multistage approach to modeling
non-stationary covariances that is efficient for large data sets. First, we use
likelihood estimation in local, moving windows to infer spatially varying
covariance parameters. These surfaces of covariance parameters can then be
encoded into a global covariance model specifying the second-order structure
for the complete spatial domain. The resulting global model allows for
efficient simulation and prediction. We investigate the non-stationary spatial
autoregressive (SAR) model related to Gaussian Markov random field (GMRF)
methods, which is amenable to plug in local estimates and practical for large
data sets. In addition we use a simulation study to establish the accuracy of
local Mat\'ern parameter estimation as a reliable technique when replicate
fields are available and small local windows are exploited to reduce
computation. This multistage modeling approach is implemented on a
non-stationary climate model output data set with the goal of emulating the
variation in the numerical model ensemble using a Gaussian process.
| stat |
Bayesian Nonparametric Boolean Factor Models | We build upon probabilistic models for Boolean Matrix and Boolean Tensor
factorisation that have recently been shown to solve these problems with
unprecedented accuracy and to enable posterior inference to scale to Billions
of observation. Here, we lift the restriction of a pre-specified number of
latent dimensions by introducing an Indian Buffet Process prior over factor
matrices. Not only does the full factor-conditional take a computationally
convenient form due to the logical dependencies in the model, but also the
posterior over the number of non-zero latent dimensions is remarkably simple.
It amounts to counting the number false and true negative predictions, whereas
positive predictions can be ignored. This constitutes a very transparent
example of sampling-based posterior inference with an IBP prior and,
importantly, lets us maintain extremely efficient inference. We discuss
applications to simulated data, as well as to a real world data matrix with 6
Million entries.
| stat |
Selecting Reduced Models in the Cross-Entropy Method | This paper deals with the estimation of rare event probabilities using
importance sampling (IS), where an optimal proposal distribution is computed
with the cross-entropy (CE) method. Although, IS optimized with the CE method
leads to an efficient reduction of the estimator variance, this approach
remains unaffordable for problems where the repeated evaluation of the score
function represents a too intensive computational effort. This is often the
case for score functions related to the solution of a partial differential
equation (PDE) with random inputs. This work proposes to alleviate computation
by the parsimonious use of a hierarchy of score function approximations in the
CE optimization process. The score function approximation is obtained by
selecting the surrogate of lowest dimensionality, whose accuracy guarantees to
pass the current CE optimization stage. The selection relies on certified upper
bounds on the error norm. An asymptotic analysis provides some theoretical
guarantees on the efficiency and convergence of the proposed algorithm.
Numerical results demonstrate the gain brought by the method in the context of
pollution alerts and a system modeled by a PDE.
| stat |
Integrated Nested Laplace Approximations (INLA) | This is a short description and basic introduction to the Integrated nested
Laplace approximations (INLA) approach. INLA is a deterministic paradigm for
Bayesian inference in latent Gaussian models (LGMs) introduced in Rue et al.
(2009). INLA relies on a combination of analytical approximations and efficient
numerical integration schemes to achieve highly accurate deterministic
approximations to posterior quantities of interest. The main benefit of using
INLA instead of Markov chain Monte Carlo (MCMC) techniques for LGMs is
computational; INLA is fast even for large, complex models. Moreover, being a
deterministic algorithm, INLA does not suffer from slow convergence and poor
mixing. INLA is implemented in the R package R-INLA, which represents a
user-friendly and versatile tool for doing Bayesian inference. R-INLA returns
posterior marginals for all model parameters and the corresponding posterior
summary information. Model choice criteria as well as predictive diagnostics
are directly available. Here, we outline the theory behind INLA, present the
R-INLA package and describe new developments of combining INLA with MCMC for
models that are not possible to fit with R-INLA.
| stat |
Model-agnostic Feature Importance and Effects with Dependent Features --
A Conditional Subgroup Approach | Partial dependence plots and permutation feature importance are popular
model-agnostic interpretation methods. Both methods are based on predicting
artificially created data points. When features are dependent, both methods
extrapolate to feature areas with low data density. The extrapolation can cause
misleading interpretations. To overcome extrapolation, we propose conditional
variants of partial dependence plots and permutation feature importance. Our
approach is based on perturbations in subgroups. The subgroups partition the
feature space to make the feature distribution within a group more homogeneous
and between the groups more heterogeneous. The interpretable subgroups enable
additional local, nuanced interpretations of the feature dependence structure
as well as the feature effects and importance values within the subgroups. We
also introduce a data fidelity measure that captures the degree of
extrapolation when data is transformed with a certain perturbation. In
simulations and benchmarks on real data we show that our conditional
interpretation methods reduce extrapolation. In an application we show that
these methods provide more nuanced and richer explanations.
| stat |
Implementation of Correlation and Regression Models for Health Insurance
Fraud in Covid-19 Environment using Actuarial and Data Science Techniques | Fraud acts as a major deterrent to a companys growth if uncontrolled. It
challenges the fundamental value of Trust in the Insurance business. COVID-19
brought additional challenges of increased potential fraud to health insurance
business. This work describes implementation of existing and enhanced fraud
detection methods in the pre-COVID-19 and COVID-19 environments. For this
purpose, we have developed an innovative enhanced fraud detection framework
using actuarial and data science techniques. Triggers specific to COVID-19 are
identified in addition to the existing triggers. We have also explored the
relationship between insurance fraud and COVID-19. To determine this we
calculated Pearson correlation coefficient and fitted logarithmic regression
model between fraud in health insurance and COVID-19 cases. This work uses two
datasets: health insurance dataset and Kaggle dataset on COVID-19 cases for the
same select geographical location in India. Our experimental results shows
Pearson correlation coefficient of 0.86, which implies that the month on month
rate of fraudulent cases is highly correlated with month on month rate of
COVID-19 cases. The logarithmic regression performed on the data gave the
r-squared value of 0.91 which indicates that the model is a good fit. This work
aims to provide much needed tools and techniques for health insurance business
to counter the fraud.
| stat |
SODEN: A Scalable Continuous-Time Survival Model through Ordinary
Differential Equation Networks | In this paper, we propose a flexible model for survival analysis using neural
networks along with scalable optimization algorithms. One key technical
challenge for directly applying maximum likelihood estimation (MLE) to censored
data is that evaluating the objective function and its gradients with respect
to model parameters requires the calculation of integrals. To address this
challenge, we recognize that the MLE for censored data can be viewed as a
differential-equation constrained optimization problem, a novel perspective.
Following this connection, we model the distribution of event time through an
ordinary differential equation and utilize efficient ODE solvers and adjoint
sensitivity analysis to numerically evaluate the likelihood and the gradients.
Using this approach, we are able to 1) provide a broad family of
continuous-time survival distributions without strong structural assumptions,
2) obtain powerful feature representations using neural networks, and 3) allow
efficient estimation of the model in large-scale applications using stochastic
gradient descent. Through both simulation studies and real-world data examples,
we demonstrate the effectiveness of the proposed method in comparison to
existing state-of-the-art deep learning survival analysis models.
| stat |
Efficiently transporting causal (in)direct effects to new populations
under intermediate confounding and with multiple mediators | The same intervention can produce different effects in different sites.
Transport mediation estimators can estimate the extent to which such
differences can be explained by differences in compositional factors and the
mechanisms by which mediating or intermediate variables are produced; however,
they are limited to consider a single, binary mediator. We propose novel
nonparametric estimators of transported stochastic (in)direct effects that
consider multiple, high-dimensional mediators and intermediate variables. They
are multiply robust, efficient, asymptotically normal, and can incorporate
data-adaptive estimation of nuisance parameters. They can be applied to
understand differences in treatment effects across sites and/or to predict
treatment effects in a target site based on outcome data in source sites.
| stat |
Streaming computation of optimal weak transport barycenters | We introduce the weak barycenter of a family of probability distributions,
based on the recently developed notion of optimal weak transport of measures
arXiv:1412.7480(v4). We provide a theoretical analysis of the weak barycenter
and its relationship to the classic Wasserstein barycenter, and discuss its
meaning in the light of convex ordering between probability measures. In
particular, we argue that, rather than averaging the information of the input
distributions as done by the usual optimal transport barycenters, weak
barycenters contain geometric information shared across all input
distributions, which can be interpreted as a latent random variable affecting
all the measures. We also provide iterative algorithms to compute a weak
barycenter for either finite or infinite families of arbitrary measures (with
finite moments of order 2), which are particularly well suited for the
streaming setting, i.e., when measures arrive sequentially. In particular, our
streaming computation of weak barycenters does not require to smooth empirical
measures or to define a common grid for them, as some of the previous
approaches to Wasserstin barycenters do. The concept of weak barycenter and our
computation approaches are illustrated on synthetic examples, validated on 2D
real-world data and compared to the classical Wasserstein barycenters.
| stat |
Bayesian Projected Calibration of Computer Models | We develop a Bayesian approach called Bayesian projected calibration to
address the problem of calibrating an imperfect computer model using
observational data from a complex physical system. The calibration parameter
and the physical system are parametrized in an identifiable fashion via
$L_2$-projection. The physical process is assigned a Gaussian process prior,
which naturally induces a prior distribution on the calibration parameter
through the $L_2$-projection constraint. The calibration parameter is estimated
through its posterior distribution, which provides a natural and non-asymptotic
way for the uncertainty quantification. We provide a rigorous large sample
justification for the proposed approach by establishing the asymptotic
normality of the posterior of the calibration parameter with the efficient
covariance matrix. In addition, two efficient computational algorithms based on
stochastic approximation are designed with theoretical guarantees. Through
extensive simulation studies and two real-world datasets analyses, we show that
the Bayesian projected calibration can accurately estimate the calibration
parameters, appropriately calibrate the computer models, and compare favorably
to alternative approaches.
| stat |
Independent Approximates enable closed-form estimation of heavy-tailed
distributions | Independent Approximates (IAs) are proven to enable a closed-form estimation
of heavy-tailed distributions with an analytical density such as the
generalized Pareto and Student's t distributions. A broader proof using
convolution of the characteristic function is described for future research.
(IAs) are selected from independent, identically distributed samples by
partitioning the samples into groups of size n and retaining the median of the
samples in those groups which have approximately equal samples. The marginal
distribution along the diagonal of equal values has a density proportional to
the nth power of the original density. This nth power density, which the IAs
approximate, has faster tail decay enabling closed-form estimation of its
moments and retains a functional relationship with the original density.
Computational experiments with between 1000 to 100,000 Student's t samples are
reported for over a range of location, scale, and shape (inverse of degree of
freedom) parameters. IA pairs are used to estimate the location, IA triplets
for the scale, and the geometric mean of the original samples for the shape.
With 10,000 samples the relative bias of the parameter estimates is less than
0.01 and a relative precision is less than plus or minus 0.1. The theoretical
bias is zero for the location and the finite bias for the scale can be
subtracted out. The theoretical precision has a finite range when the shape is
less than 2 for the location estimate and less than 3/2 for the scale estimate.
The boundary of finite precision can be extended using higher-order IAs.
| stat |
Identifying overlapping terrorist cells from the Noordin Top actor-event
network | Actor-event data are common in sociological settings, whereby one registers
the pattern of attendance of a group of social actors to a number of events. We
focus on 79 members of the Noordin Top terrorist network, who were monitored
attending 45 events. The attendance or non-attendance of the terrorist to
events defines the social fabric, such as group coherence and social
communities. The aim of the analysis of such data is to learn about the
affiliation structure. Actor-event data is often transformed to actor-actor
data in order to be further analysed by network models, such as stochastic
block models. This transformation and such analyses lead to a natural loss of
information, particularly when one is interested in identifying, possibly
overlapping, subgroups or communities of actors on the basis of their
attendances to events. In this paper we propose an actor-event model for
overlapping communities of terrorists, which simplifies interpretation of the
network. We propose a mixture model with overlapping clusters for the analysis
of the binary actor-event network data, called {\tt manet}, and develop a
Bayesian procedure for inference. After a simulation study, we show how this
analysis of the terrorist network has clear interpretative advantages over the
more traditional approaches of affiliation network analysis.
| stat |
Sparse Identification of Nonlinear Dynamical Systems via Reweighted
$\ell_1$-regularized Least Squares | This work proposes an iterative sparse-regularized regression method to
recover governing equations of nonlinear dynamical systems from noisy state
measurements. The method is inspired by the Sparse Identification of Nonlinear
Dynamics (SINDy) approach of {\it [Brunton et al., PNAS, 113 (15) (2016)
3932-3937]}, which relies on two main assumptions: the state variables are
known {\it a priori} and the governing equations lend themselves to sparse,
linear expansions in a (nonlinear) basis of the state variables. The aim of
this work is to improve the accuracy and robustness of SINDy in the presence of
state measurement noise. To this end, a reweighted $\ell_1$-regularized least
squares solver is developed, wherein the regularization parameter is selected
from the corner point of a Pareto curve. The idea behind using weighted
$\ell_1$-norm for regularization -- instead of the standard $\ell_1$-norm -- is
to better promote sparsity in the recovery of the governing equations and, in
turn, mitigate the effect of noise in the state variables. We also present a
method to recover single physical constraints from state measurements. Through
several examples of well-known nonlinear dynamical systems, we demonstrate
empirically the accuracy and robustness of the reweighted $\ell_1$-regularized
least squares strategy with respect to state measurement noise, thus
illustrating its viability for a wide range of potential applications.
| stat |
Learning to Project in Multi-Objective Binary Linear Programming | In this paper, we investigate the possibility of improving the performance of
multi-objective optimization solution approaches using machine learning
techniques. Specifically, we focus on multi-objective binary linear programs
and employ one of the most effective and recently developed criterion space
search algorithms, the so-called KSA, during our study. This algorithm computes
all nondominated points of a problem with p objectives by searching on a
projected criterion space, i.e., a (p-1)-dimensional criterion apace. We
present an effective and fast learning approach to identify on which projected
space the KSA should work. We also present several generic features/variables
that can be used in machine learning techniques for identifying the best
projected space. Finally, we present an effective bi-objective optimization
based heuristic for selecting the best subset of the features to overcome the
issue of overfitting in learning. Through an extensive computational study over
2000 instances of tri-objective Knapsack and Assignment problems, we
demonstrate that an improvement of up to 12% in time can be achieved by the
proposed learning method compared to a random selection of the projected space.
| stat |
Algorithms for ridge estimation with convergence guarantees | The extraction of filamentary structure from a point cloud is discussed. The
filaments are modeled as ridge lines or higher dimensional ridges of an
underlying density. We propose two novel algorithms, and provide theoretical
guarantees for their convergences. We consider the new algorithms as
alternatives to the Subspace Constraint Mean Shift (SCMS) algorithm that do not
suffer from a shortcoming of the SCMS that is also revealed in this paper.
| stat |
Estimating Knots and Their Association in Parallel Bilinear Spline
Growth Curve Models in the Framework of Individual Measurement Occasions | Latent growth curve models with spline functions are flexible and accessible
statistical tools for investigating nonlinear change patterns that exhibit
distinct phases of development in manifested variables. Among such models, the
bilinear spline growth model (BLSGM) is the most straightforward and intuitive
but useful. An existing study has demonstrated that the BLSGM allows the knot
(or change-point), at which two linear segments join together, to be an
additional growth factor other than the intercept and slopes so that
researchers can estimate the knot and its variability in the framework of
individual measurement occasions. However, developmental processes usually
unfold in a joint development where two or more outcomes and their change
patterns are correlated over time. As an extension of the existing BLSGM with
an unknown knot, this study considers a parallel BLSGM (PBLSGM) for
investigating multiple nonlinear growth processes and estimating the knot with
its variability of each process as well as the knot-knot association in the
framework of individual measurement occasions. We present the proposed model by
simulation studies and a real-world data analysis. Our simulation studies
demonstrate that the proposed PBLSGM generally estimate the parameters of
interest unbiasedly, precisely and exhibit appropriate confidence interval
coverage. An empirical example using longitudinal reading scores, mathematics
scores, and science scores shows that the model can estimate the knot with its
variance for each growth curve and the covariance between two knots. We also
provide the corresponding code for the proposed model.
| stat |
Probabilistic modeling of discrete structural response with application
to composite plate penetration models | Discrete response of structures is often a key probabilistic quantity of
interest. For example, one may need to identify the probability of a binary
event, such as, whether a structure has buckled or not. In this study, an
adaptive domain-based decomposition and classification method, combined with
sparse grid sampling, is used to develop an efficient classification surrogate
modeling algorithm for such discrete outputs. An assumption of monotonic
behaviour of the output with respect to all model parameters, based on the
physics of the problem, helps to reduce the number of model evaluations and
makes the algorithm more efficient. As an application problem, this paper deals
with the development of a computational framework for generation of
probabilistic penetration response of S-2 glass/SC-15 epoxy composite plates
under ballistic impact. This enables the computationally feasible generation of
the probabilistic velocity response (PVR) curve or the $V_0-V_{100}$ curve as a
function of the impact velocity, and the ballistic limit velocity prediction as
a function of the model parameters. The PVR curve incorporates the variability
of the model input parameters and describes the probability of penetration of
the plate as a function of impact velocity.
| stat |
Bayesian model selection for unsupervised image deconvolution with
structured Gaussian priors | This paper considers the objective comparison of stochastic models to solve
inverse problems, more specifically image restoration. Most often, model
comparison is addressed in a supervised manner, that can be time-consuming and
partly arbitrary. Here we adopt an unsupervised Bayesian approach and
objectively compare the models based on their posterior probabilities, directly
from the data without ground truth available. The probabilities depend on the
marginal likelihood or "evidence" of the models and we resort to the Chib
approach including a Gibbs sampler. We focus on the family of Gaussian models
with circulant covariances and unknown hyperparameters, and compare different
types of covariance matrices for the image and noise.
| stat |
Infectious Disease Forecasting for Public Health | Forecasting transmission of infectious diseases, especially for vector-borne
diseases, poses unique challenges for researchers. Behaviors of and
interactions between viruses, vectors, hosts, and the environment each play a
part in determining the transmission of a disease. Public health surveillance
systems and other sources provide valuable data that can be used to accurately
forecast disease incidence. However, many aspects of common infectious disease
surveillance data are imperfect: cases may be reported with a delay or in some
cases not at all, data on vectors may not be available, and case data may not
be available at high geographical or temporal resolution. In the face of these
challenges, researchers must make assumptions to either account for these
underlying processes in a mechanistic model or to justify their exclusion
altogether in a statistical model. Whether a model is mechanistic or
statistical, researchers should evaluate their model using accepted best
practices from the emerging field of infectious disease forecasting while
adopting conventions from other fields that have been developing forecasting
methods for decades. Accounting for assumptions and properly evaluating models
will allow researchers to generate forecasts that have the potential to provide
valuable insights for public health officials. This chapter provides a
background to the practice of forecasting in general, discusses the biological
and statistical models used for infectious disease forecasting, presents
technical details about making and evaluating forecasting models, and explores
the issues in communicating forecasting results in a public health context.
| stat |
Dimension Independent Generalization Error by Stochastic Gradient
Descent | One classical canon of statistics is that large models are prone to
overfitting, and model selection procedures are necessary for high dimensional
data. However, many overparameterized models, such as neural networks, perform
very well in practice, although they are often trained with simple online
methods and regularization. The empirical success of overparameterized models,
which is often known as benign overfitting, motivates us to have a new look at
the statistical generalization theory for online optimization. In particular,
we present a general theory on the generalization error of stochastic gradient
descent (SGD) solutions for both convex and locally convex loss functions. We
further discuss data and model conditions that lead to a ``low effective
dimension". Under these conditions, we show that the generalization error
either does not depend on the ambient dimension $p$ or depends on $p$ via a
poly-logarithmic factor. We also demonstrate that in several widely used
statistical models, the ``low effective dimension'' arises naturally in
overparameterized settings. The studied statistical applications include both
convex models such as linear regression and logistic regression and non-convex
models such as $M$-estimator and two-layer neural networks.
| stat |
Partial Separability and Functional Graphical Models for Multivariate
Gaussian Processes | The covariance structure of multivariate functional data can be highly
complex, especially if the multivariate dimension is large, making extension of
statistical methods for standard multivariate data to the functional data
setting quite challenging. For example, Gaussian graphical models have recently
been extended to the setting of multivariate functional data by applying
multivariate methods to the coefficients of truncated basis expansions.
However, a key difficulty compared to multivariate data is that the covariance
operator is compact, and thus not invertible. The methodology in this paper
addresses the general problem of covariance modeling for multivariate
functional data, and functional Gaussian graphical models in particular. As a
first step, a new notion of separability for multivariate functional data is
proposed, termed partial separability, leading to a novel Karhunen-Lo\`eve-type
expansion for such data. Next, the partial separability structure is shown to
be particularly useful in order to provide a well-defined Gaussian graphical
model that can be identified with a sequence of finite-dimensional graphical
models, each of fixed dimension. This motivates a simple and efficient
estimation procedure through application of the joint graphical lasso.
Empirical performance of the method for graphical model estimation is assessed
through simulation and analysis of functional brain connectivity during a motor
task.
| stat |
Instrumental Variable Estimation of Marginal Structural Mean Models for
Time-Varying Treatment | Robins 1997 introduced marginal structural models (MSMs), a general class of
counterfactual models for the joint effects of time-varying treatment regimes
in complex longitudinal studies subject to time-varying confounding. In his
work, identification of MSM parameters is established under a sequential
randomization assumption (SRA), which rules out unmeasured confounding of
treatment assignment over time. We consider sufficient conditions for
identification of the parameters of a subclass, Marginal Structural Mean Models
(MSMMs), when sequential randomization fails to hold due to unmeasured
confounding, using instead a time-varying instrumental variable. Our
identification conditions require that no unobserved confounder predicts
compliance type for the time-varying treatment. We describe a simple weighted
estimator and examine its finite-sample properties in a simulation study. We
apply the proposed estimator to examine the effect of delivery hospital on
neonatal survival probability.
| stat |
Multivariate Time-Between-Events Monitoring -- An overview and some
(overlooked) underlying complexities | We review methods for monitoring multivariate time-between-events (TBE) data.
We present some underlying complexities that have been overlooked in the
literature. It is helpful to classify multivariate TBE monitoring applications
into two fundamentally different scenarios. One scenario involves monitoring
individual vectors of TBE data. The other involves the monitoring of several,
possibly correlated, temporal point processes in which events could occur at
different rates. We discuss performance measures and advise the use of
time-between-signal based metrics for the design and comparison of methods. We
re-evaluate an existing multivariate TBE monitoring method, offer some advice
and some directions for future research.
| stat |
Algorithms for finding generalized minimum aberration designs | Statistical design of experiments is widely used in scientific and industrial
investigations. A generalized minimum aberration (GMA) orthogonal array is
optimum under the well-established, so-called GMA criterion, and such an array
can extract as much information as possible at a fixed cost. Finding GMA arrays
is an open (yet fundamental) problem in design of experiments because
constructing such arrays becomes intractable as the number of runs and factors
increase. We develop two directed enumeration algorithms that call the integer
programming with isomorphism pruning algorithm of Margot (2007) for the purpose
of finding GMA arrays. Our results include 16 GMA arrays that were not
previously in the literature, along with documentation of the efficiencies that
made the required calculations possible within a reasonable budget of computer
time. We also validate heuristic algorithms against a GMA array catalog, by
showing that they quickly output near GMA arrays, and then use the heuristics
to find near GMA arrays when enumeration is computationally burdensome.
| stat |
Distribution-Free Pointwise Adjusted P-Values for Functional Hypotheses | Graphical tests assess whether a function of interest departs from an
envelope of functions generated under a simulated null distribution. This
approach originated in spatial statistics, but has recently gained some
popularity in functional data analysis. Whereas such envelope tests examine
deviation from a functional null distribution in an omnibus sense, in some
applications we wish to do more: to obtain p-values at each point in the
function domain, adjusted to control the familywise error rate. Here we derive
pointwise adjusted p-values based on envelope tests, and relate these to
previous approaches for functional data under distributional assumptions. We
then present two alternative distribution-free p-value adjustments that offer
greater power. The methods are illustrated with an analysis of age-varying sex
effects on cortical thickness in the human brain.
| stat |
The Neural Tangent Kernel in High Dimensions: Triple Descent and a
Multi-Scale Theory of Generalization | Modern deep learning models employ considerably more parameters than required
to fit the training data. Whereas conventional statistical wisdom suggests such
models should drastically overfit, in practice these models generalize
remarkably well. An emerging paradigm for describing this unexpected behavior
is in terms of a \emph{double descent} curve, in which increasing a model's
capacity causes its test error to first decrease, then increase to a maximum
near the interpolation threshold, and then decrease again in the
overparameterized regime. Recent efforts to explain this phenomenon
theoretically have focused on simple settings, such as linear regression or
kernel regression with unstructured random features, which we argue are too
coarse to reveal important nuances of actual neural networks. We provide a
precise high-dimensional asymptotic analysis of generalization under kernel
regression with the Neural Tangent Kernel, which characterizes the behavior of
wide neural networks optimized with gradient descent. Our results reveal that
the test error has non-monotonic behavior deep in the overparameterized regime
and can even exhibit additional peaks and descents when the number of
parameters scales quadratically with the dataset size.
| stat |
Causal isotonic regression | In observational studies, potential confounders may distort the causal
relationship between an exposure and an outcome. However, under some
conditions, a causal dose-response curve can be recovered using the
G-computation formula. Most classical methods for estimating such curves when
the exposure is continuous rely on restrictive parametric assumptions, which
carry significant risk of model misspecification. Nonparametric estimation in
this context is challenging because in a nonparametric model these curves
cannot be estimated at regular rates. Many available nonparametric estimators
are sensitive to the selection of certain tuning parameters, and performing
valid inference with such estimators can be difficult. In this work, we propose
a nonparametric estimator of a causal dose-response curve known to be monotone.
We show that our proposed estimation procedure generalizes the classical
least-squares isotonic regression estimator of a monotone regression function.
Specifically, it does not involve tuning parameters, and is invariant to
strictly monotone transformations of the exposure variable. We describe
theoretical properties of our proposed estimator, including its irregular limit
distribution and the potential for doubly-robust inference. Furthermore, we
illustrate its performance via numerical studies, and use it to assess the
relationship between BMI and immune response in HIV vaccine trials.
| stat |
Bayesian Survival Analysis Using the rstanarm R Package | Survival data is encountered in a range of disciplines, most notably health
and medical research. Although Bayesian approaches to the analysis of survival
data can provide a number of benefits, they are less widely used than classical
(e.g. likelihood-based) approaches. This may be in part due to a relative
absence of user-friendly implementations of Bayesian survival models. In this
article we describe how the rstanarm R package can be used to fit a wide range
of Bayesian survival models. The rstanarm package facilitates Bayesian
regression modelling by providing a user-friendly interface (users specify
their model using customary R formula syntax and data frames) and using the
Stan software (a C++ library for Bayesian inference) for the back-end
estimation. The suite of models that can be estimated using rstanarm is broad
and includes generalised linear models (GLMs), generalised linear mixed models
(GLMMs), generalised additive models (GAMs) and more. In this article we focus
only on the survival modelling functionality. This includes standard parametric
(exponential, Weibull, Gompertz) and flexible parametric (spline-based) hazard
models, as well as standard parametric accelerated failure time (AFT) models.
All types of censoring (left, right, interval) are allowed, as is delayed entry
(left truncation), time-varying covariates, time-varying effects, and frailty
effects. We demonstrate the functionality through worked examples. We
anticipate these implementations will increase the uptake of Bayesian survival
analysis in applied research.
| stat |
Deciding with Judgment | A decision maker starts from a judgmental decision and moves to the closest
boundary of the confidence interval. This statistical decision rule is
admissible and does not perform worse than the judgmental decision with a
probability equal to the confidence level, which is interpreted as a
coefficient of statistical risk aversion. The confidence level is related to
the decision maker's aversion to uncertainty and can be elicited with
laboratory experiments using urns a la Ellsberg. The decision rule is applied
to a problem of asset allocation for an investor whose judgmental decision is
to keep all her wealth in cash.
| stat |
Statistical Inference for Qualitative Interactions with Applications to
Precision Medicine and Differential Network Analysis | Qualitative interactions occur when a treatment effect or measure of
association varies in sign by sub-population. Of particular interest in many
biomedical settings are absence/presence qualitative interactions, which occur
when an effect is present in one sub-population but absent in another.
Absence/presence interactions arise in emerging applications in precision
medicine, where the objective is to identify a set of predictive biomarkers
that have prognostic value for clinical outcomes in some sub-population but not
others. They also arise naturally in gene regulatory network inference, where
the goal is to identify differences in networks corresponding to diseased and
healthy individuals, or to different subtypes of disease; such differences lead
to identification of network-based biomarkers for diseases. In this paper, we
argue that while the absence/presence hypothesis is important, developing a
statistical test for this hypothesis is an intractable problem. To overcome
this challenge, we approximate the problem in a novel inference framework. In
particular, we propose to make inferences about absence/presence interactions
by quantifying the relative difference in effect size, reasoning that when the
relative difference is large, an absence/presence interaction occurs. The
proposed methodology is illustrated through a simulation study as well as an
analysis of breast cancer data from the Cancer Genome Atlas.
| stat |
Large-data determinantal clustering | Determinantal consensus clustering is a promising and attractive alternative
to partitioning about medoids and k-means for ensemble clustering. Based on a
determinantal point process or DPP sampling, it ensures that subsets of similar
points are less likely to be selected as centroids. It favors more diverse
subsets of points. The sampling algorithm of the determinantal point process
requires the eigendecomposition of a Gram matrix. This becomes computationally
intensive when the data size is very large. This is particularly an issue in
consensus clustering, where a given clustering algorithm is run several times
in order to produce a final consolidated clustering. We propose two efficient
alternatives to carry out determinantal consensus clustering on large datasets.
They consist in DPP sampling based on sparse and small kernel matrices whose
eigenvalue distributions are close to that of the original Gram matrix.
| stat |
Inference after black box selection | We consider the problem of inference for parameters selected to report only
after some algorithm, the canonical example being inference for model
parameters after a model selection procedure. The conditional correction for
selection requires knowledge of how the selection is affected by changes in the
underlying data, and current research explicitly describes this selection. In
this work, we assume 1) we have in silico access to the selection algorithm and
2) for parameters of interest, the data input into the algorithm satisfies
(pre-selection) a central limit theorem jointly with an estimator of our
parameter of interest. Under these assumptions, we recast the problem into a
statistical learning problem which can be fit with off-the-shelf models for
binary regression. The feature points in this problem are set by the user,
opening up the possibility of active learning methods for computationally
expensive selection algorithms. We consider two examples previously out of
reach of this conditional approach: stability selection and multiple
cross-validation.
| stat |
Fragmentation Coagulation Based Mixed Membership Stochastic Blockmodel | The Mixed-Membership Stochastic Blockmodel~(MMSB) is proposed as one of the
state-of-the-art Bayesian relational methods suitable for learning the complex
hidden structure underlying the network data. However, the current formulation
of MMSB suffers from the following two issues: (1), the prior information~(e.g.
entities' community structural information) can not be well embedded in the
modelling; (2), community evolution can not be well described in the
literature. Therefore, we propose a non-parametric fragmentation coagulation
based Mixed Membership Stochastic Blockmodel (fcMMSB). Our model performs
entity-based clustering to capture the community information for entities and
linkage-based clustering to derive the group information for links
simultaneously. Besides, the proposed model infers the network structure and
models community evolution, manifested by appearances and disappearances of
communities, using the discrete fragmentation coagulation process (DFCP). By
integrating the community structure with the group compatibility matrix we
derive a generalized version of MMSB. An efficient Gibbs sampling scheme with
Polya Gamma (PG) approach is implemented for posterior inference. We validate
our model on synthetic and real world data.
| stat |
Lung Cancer Detection and Classification based on Image Processing and
Statistical Learning | Lung cancer is one of the death threatening diseases among human beings.
Early and accurate detection of lung cancer can increase the survival rate from
lung cancer. Computed Tomography (CT) images are commonly used for detecting
the lung cancer.Using a data set of thousands of high-resolution lung scans
collected from Kaggle competition [1], we will develop algorithms that
accurately determine in the lungs are cancerous or not. The proposed system
promises better result than the existing systems, which would be beneficial for
the radiologist for the accurate and early detection of cancer. The method has
been tested on 198 slices of CT images of various stages of cancer obtained
from Kaggle dataset[1] and is found satisfactory results. The accuracy of the
proposed method in this dataset is 72.2%
| stat |
Inference for Individual Mediation Effects and Interventional Effects in
Sparse High-Dimensional Causal Graphical Models | We consider the problem of identifying intermediate variables (or mediators)
that regulate the effect of a treatment on a response variable. While there has
been significant research on this classical topic, little work has been done
when the set of potential mediators is high-dimensional (HD). A further
complication arises when these mediators are interrelated (with unknown
dependencies). In particular, we assume that the causal structure of the
treatment, the confounders, the potential mediators and the response is a
(possibly unknown) directed acyclic graph (DAG). HD DAG models have previously
been used for the estimation of causal effects from observational data. In
particular, methods called IDA and joint-IDA have been developed for estimating
the effects of single and multiple simultaneous interventions, respectively. In
this paper, we propose an IDA-type method called MIDA for estimating so-called
individual mediation effects from HD observational data. Although IDA and
joint-IDA estimators have been shown to be consistent in certain sparse HD
settings, their asymptotic properties such as convergence in distribution and
inferential tools in such settings have remained unknown. In this paper, we
prove HD consistency of MIDA for linear structural equation models with
sub-Gaussian errors. More importantly, we derive distributional convergence
results for MIDA in similar HD settings, which are applicable to IDA and
joint-IDA estimators as well. To our knowledge, these are the first such
distributional convergence results facilitating inference for IDA-type
estimators. These are built on our novel theoretical results regarding uniform
bounds for linear regression estimators over varying subsets of HD covariates
which may be of independent interest. Finally, we empirically validate our
asymptotic theory for MIDA and demonstrate its usefulness via simulations and a
real data application.
| stat |
Minimizing Negative Transfer of Knowledge in Multivariate Gaussian
Processes: A Scalable and Regularized Approach | Recently there has been an increasing interest in the multivariate Gaussian
process (MGP) which extends the Gaussian process (GP) to deal with multiple
outputs. One approach to construct the MGP and account for non-trivial
commonalities amongst outputs employs a convolution process (CP). The CP is
based on the idea of sharing latent functions across several convolutions.
Despite the elegance of the CP construction, it provides new challenges that
need yet to be tackled. First, even with a moderate number of outputs, model
building is extremely prohibitive due to the huge increase in computational
demands and number of parameters to be estimated. Second, the negative transfer
of knowledge may occur when some outputs do not share commonalities. In this
paper we address these issues. We propose a regularized pairwise modeling
approach for the MGP established using CP. The key feature of our approach is
to distribute the estimation of the full multivariate model into a group of
bivariate GPs which are individually built. Interestingly pairwise modeling
turns out to possess unique characteristics, which allows us to tackle the
challenge of negative transfer through penalizing the latent function that
facilitates information sharing in each bivariate model. Predictions are then
made through combining predictions from the bivariate models within a Bayesian
framework. The proposed method has excellent scalability when the number of
outputs is large and minimizes the negative transfer of knowledge between
uncorrelated outputs. Statistical guarantees for the proposed method are
studied and its advantageous features are demonstrated through numerical
studies.
| stat |
On application of a response propensity model to estimation from web
samples | Increasing nonresponse rates and the cost of data collection are two pressing
problems encountered in traditional probability surveys. The proliferation of
inexpensive data from web surveys stimulates interest in statistical techniques
for valid inferences from web samples. We consider estimation of population and
domain means in the two-sample setup, where the web sample contains variables
of interest and covariates that are shared with an auxiliary probability survey
sample. First, we propose an estimator of population mean, based on the
estimated propensity of response to a web survey. This makes inferences from
web samples that are similar to well-established techniques used for
observational studies and missing data problems. Second, we propose an
'implicit' logistic regression for estimating parameters of the web response
model in the two-sample setup. Implicit logistic regression uses selection
probabilities, nominally defined for web sample units, and the size of the
hypothetic population of responders to a web survey. A simulation study
confirms the validity of implicit logistic regression and its higher efficiency
comparing to alternative estimators of web response propensity.
| stat |
A feature-based framework for detecting technical outliers in
water-quality data from in situ sensors | Outliers due to technical errors in water-quality data from in situ sensors
can reduce data quality and have a direct impact on inference drawn from
subsequent data analysis. However, outlier detection through manual monitoring
is unfeasible given the volume and velocity of data the sensors produce. Here,
we proposed an automated framework that provides early detection of outliers in
water-quality data from in situ sensors caused by technical issues.The
framework was used first to identify the data features that differentiate
outlying instances from typical behaviours. Then statistical transformations
were applied to make the outlying instances stand out in transformed data
space. Unsupervised outlier scoring techniques were then applied to the
transformed data space and an approach based on extreme value theory was used
to calculate a threshold for each potential outlier. Using two data sets
obtained from in situ sensors in rivers flowing into the Great Barrier Reef
lagoon, Australia, we showed that the proposed framework successfully
identified outliers involving abrupt changes in turbidity, conductivity and
river level, including sudden spikes, sudden isolated drops and level shifts,
while maintaining very low false detection rates. We implemented this framework
in the open source R package oddwater.
| stat |
An Evaluation of Bounding Approaches for Generalization | Statisticians have recently developed propensity score methods to improve
generalizations from randomized experiments that do not employ random sampling.
However, these methods typically rely on assumptions whose plausibility may be
questionable in practice. In this article, we introduce and discuss bounding,
an approach that is based on alternative assumptions that may be more plausible
in a given study. The bounding framework nonparametrically estimates population
parameters using a range of plausible values that are consistent with the
observed characteristics of the data. We illustrate how the bounds can be
tightened using three approaches: imposing an alternative assumption based on
monotonicity, redefining the population of inference, and using propensity
score stratification. Using the results from two simulation studies, we examine
the conditions under which bounds for the population parameter are tightened.
We conclude with an application of bounding to SimCalc, a cluster randomized
trial that evaluated the effectiveness of a technology aid on mathematics
achievement.
| stat |
Asymptotically Optimal Information-Directed Sampling | We introduce a simple and efficient algorithm for stochastic linear bandits
with finitely many actions that is asymptotically optimal and worst-case rate
optimal in finite time. The approach is based on the frequentist
information-directed sampling (IDS) framework, with a surrogate for the
information gain that is informed by the optimization problem that defines the
asymptotic lower bound. Our analysis sheds light on how IDS balances the
trade-off between regret and information. Moreover, we uncover a surprising
connection between the recently proposed primal-dual methods and the Bayesian
IDS algorithm. We demonstrate empirically that IDS is competitive with UCB in
finite-time, and can be significantly better in the asymptotic regime.
| stat |
Out of Distribution Generalization in Machine Learning | Machine learning has achieved tremendous success in a variety of domains in
recent years. However, a lot of these success stories have been in places where
the training and the testing distributions are extremely similar to each other.
In everyday situations when models are tested in slightly different data than
they were trained on, ML algorithms can fail spectacularly. This research
attempts to formally define this problem, what sets of assumptions are
reasonable to make in our data and what kind of guarantees we hope to obtain
from them. Then, we focus on a certain class of out of distribution problems,
their assumptions, and introduce simple algorithms that follow from these
assumptions that are able to provide more reliable generalization. A central
topic in the thesis is the strong link between discovering the causal structure
of the data, finding features that are reliable (when using them to predict)
regardless of their context, and out of distribution generalization.
| stat |
Recovery of sparse linear classifiers from mixture of responses | In the problem of learning a mixture of linear classifiers, the aim is to
learn a collection of hyperplanes from a sequence of binary responses. Each
response is a result of querying with a vector and indicates the side of a
randomly chosen hyperplane from the collection the query vector belongs to.
This model provides a rich representation of heterogeneous data with
categorical labels and has only been studied in some special settings. We look
at a hitherto unstudied problem of query complexity upper bound of recovering
all the hyperplanes, especially for the case when the hyperplanes are sparse.
This setting is a natural generalization of the extreme quantization problem
known as 1-bit compressed sensing. Suppose we have a set of $\ell$ unknown
$k$-sparse vectors. We can query the set with another vector $\boldsymbol{a}$,
to obtain the sign of the inner product of $\boldsymbol{a}$ and a randomly
chosen vector from the $\ell$-set. How many queries are sufficient to identify
all the $\ell$ unknown vectors? This question is significantly more challenging
than both the basic 1-bit compressed sensing problem (i.e., $\ell=1$ case) and
the analogous regression problem (where the value instead of the sign is
provided). We provide rigorous query complexity results (with efficient
algorithms) for this problem.
| stat |
Judging the Judges: A General Framework for Evaluating the Performance
of International Sports Judges | The monitoring of judges and referees in sports has become an important topic
due to the increasing media exposure of international sporting events and the
large monetary sums involved. In this article, we present a method to assess
the accuracy of sports judges and estimate their bias. Our method is broadly
applicable to all sports where panels of judges evaluate athletic performances
on a finite scale. We analyze judging scores from eight different sports with
comparable judging systems: diving, dressage, figure skating, freestyle skiing
(aerials), freestyle snowboard (halfpipe, slopestyle), gymnastics, ski jumping
and synchronized swimming. With the notable exception of dressage, we identify,
for each aforementioned sport, a general and accurate pattern of the intrinsic
judging error as a function of the performance level of the athlete. This
intrinsic judging inaccuracy is heteroscedastic and can be approximated by a
quadratic curve, indicating increased consensus among judges towards the best
athletes. Using this observation, the framework developed to assess the
performance of international gymnastics judges is applicable to all these
sports: we can evaluate the performance of judges compared to their peers and
distinguish cheating from unintentional misjudging. Our analysis also leads to
valuable insights about the judging practices of the sports under
consideration. In particular, it reveals a systemic judging problem in
dressage, where judges disagree on what constitutes a good performance.
| stat |
The e-value: A Fully Bayesian Significance Measure for Precise
Statistical Hypotheses and its Research Program | This article gives a survey of the e-value, a statistical significance
measure a.k.a. the evidence rendered by observational data, X, in support of a
statistical hypothesis, H, or, the other way around, the epistemic value of H
given X. The $e$-value and the accompanying FBST, the Full Bayesian
Significance Test, constitute the core of a research program that was started
at IME-USP, is being developed by over 20 researchers worldwide, and has, so
far, been referenced by over 200 publications.
The e-value and the FBST comply with the best principles of Bayesian
inference, including the likelihood principle, complete invariance, asymptotic
consistency, etc. Furthermore, they exhibit powerful logic or algebraic
properties in situations where one needs to compare or compose distinct
hypotheses that can be formulated either in the same or in different
statistical models. Moreover, they effortlessly accommodate the case of sharp
or precise hypotheses, a situation where alternative methods often require ad
hoc and convoluted procedures. Finally, the FBST has outstanding robustness and
reliability characteristics, outperforming traditional tests of hypotheses in
many practical applications of statistical modeling and operations research.
| stat |
Divide, Conquer, and Combine: a New Inference Strategy for Probabilistic
Programs with Stochastic Support | Universal probabilistic programming systems (PPSs) provide a powerful
framework for specifying rich probabilistic models. They further attempt to
automate the process of drawing inferences from these models, but doing this
successfully is severely hampered by the wide range of non--standard models
they can express. As a result, although one can specify complex models in a
universal PPS, the provided inference engines often fall far short of what is
required. In particular, we show that they produce surprisingly unsatisfactory
performance for models where the support varies between executions, often doing
no better than importance sampling from the prior. To address this, we
introduce a new inference framework: Divide, Conquer, and Combine, which
remains efficient for such models, and show how it can be implemented as an
automated and generic PPS inference engine. We empirically demonstrate
substantial performance improvements over existing approaches on three
examples.
| stat |
A linear time method for the detection of point and collective anomalies | The challenge of efficiently identifying anomalies in data sequences is an
important statistical problem that now arises in many applications. Whilst
there has been substantial work aimed at making statistical analyses robust to
outliers, or point anomalies, there has been much less work on detecting
anomalous segments, or collective anomalies, particularly in those settings
where point anomalies might also occur. In this article, we introduce
Collective And Point Anomalies (CAPA), a computationally efficient approach
that is suitable when collective anomalies are characterised by either a change
in mean, variance, or both, and distinguishes them from point anomalies.
Theoretical results establish the consistency of CAPA at detecting collective
anomalies and, as a by-product, the consistency of a popular penalised cost
based change in mean and variance detection method. Empirical results show that
CAPA has close to linear computational cost as well as being more accurate at
detecting and locating collective anomalies than other approaches. We
demonstrate the utility of CAPA through its ability to detect exoplanets from
light curve data from the Kepler telescope.
| stat |
Coupling the reduced-order model and the generative model for an
importance sampling estimator | In this work, we develop an importance sampling estimator by coupling the
reduced-order model and the generative model in a problem setting of
uncertainty quantification. The target is to estimate the probability that the
quantity of interest (QoI) in a complex system is beyond a given threshold. To
avoid the prohibitive cost of sampling a large scale system, the reduced-order
model is usually considered for a trade-off between efficiency and accuracy.
However, the Monte Carlo estimator given by the reduced-order model is biased
due to the error from dimension reduction. To correct the bias, we still need
to sample the fine model. An effective technique to reduce the variance
reduction is importance sampling, where we employ the generative model to
estimate the distribution of the data from the reduced-order model and use it
for the change of measure in the importance sampling estimator. To compensate
the approximation errors of the reduced-order model, more data that induce a
slightly smaller QoI than the threshold need to be included into the training
set. Although the amount of these data can be controlled by a posterior error
estimate, redundant data, which may outnumber the effective data, will be kept
due to the epistemic uncertainty. To deal with this issue, we introduce a
weighted empirical distribution to process the data from the reduced-order
model. The generative model is then trained by minimizing the cross entropy
between it and the weighted empirical distribution. We also introduce a penalty
term into the objective function to deal with the overfitting for more
robustness. Numerical results are presented to demonstrate the effectiveness of
the proposed methodology.
| stat |
Automatic Tuning of Stochastic Gradient Descent with Bayesian
Optimisation | Many machine learning models require a training procedure based on running
stochastic gradient descent. A key element for the efficiency of those
algorithms is the choice of the learning rate schedule. While finding good
learning rates schedules using Bayesian optimisation has been tackled by
several authors, adapting it dynamically in a data-driven way is an open
question. This is of high practical importance to users that need to train a
single, expensive model. To tackle this problem, we introduce an original
probabilistic model for traces of optimisers, based on latent Gaussian
processes and an auto-/regressive formulation, that flexibly adjusts to abrupt
changes of behaviours induced by new learning rate values. As illustrated, this
model is well-suited to tackle a set of problems: first, for the on-line
adaptation of the learning rate for a cold-started run; then, for tuning the
schedule for a set of similar tasks (in a classical BO setup), as well as
warm-starting it for a new task.
| stat |
A Causal Inference Approach to Measure the Vulnerability of Urban Metro
Systems | Transit operators need vulnerability measures to understand the level of
service degradation under disruptions. This paper contributes to the literature
with a novel causal inference approach for estimating station-level
vulnerability in metro systems. The empirical analysis is based on large-scale
data on historical incidents and population-level passenger demand. This
analysis thus obviates the need for assumptions made by previous studies on
human behaviour and disruption scenarios. We develop four empirical
vulnerability metrics based on the causal impact of disruptions on travel
demand, average travel speed and passenger flow distribution. Specifically, the
proposed metrics based on the irregularity in passenger flow distribution
extends the scope of vulnerability measurement to the entire trip distribution,
instead of just analysing the disruption impact on the entry or exit demand
(that is, moments of the trip distribution). The unbiased estimates of
disruption impact are obtained by adopting a propensity score matching method,
which adjusts for the confounding biases caused by non-random occurrence of
disruptions. An application of the proposed framework to the London Underground
indicates that the vulnerability of a metro station depends on the location,
topology, and other characteristics. We find that, in 2013, central London
stations are more vulnerable in terms of travel demand loss. However, the loss
of average travel speed and irregularity in relative passenger flows reveal
that passengers from outer London stations suffer from longer individual delays
due to lack of alternative routes.
| stat |
Concept Tree: High-Level Representation of Variables for More
Interpretable Surrogate Decision Trees | Interpretable surrogates of black-box predictors trained on high-dimensional
tabular datasets can struggle to generate comprehensible explanations in the
presence of correlated variables. We propose a model-agnostic interpretable
surrogate that provides global and local explanations of black-box classifiers
to address this issue. We introduce the idea of concepts as intuitive groupings
of variables that are either defined by a domain expert or automatically
discovered using correlation coefficients. Concepts are embedded in a surrogate
decision tree to enhance its comprehensibility. First experiments on FRED-MD, a
macroeconomic database with 134 variables, show improvement in
human-interpretability while accuracy and fidelity of the surrogate model are
preserved.
| stat |
Heterogeneous causal effects with imperfect compliance: a novel Bayesian
machine learning approach | This paper introduces an innovative Bayesian machine learning algorithm to
draw interpretable inference on heterogeneous causal effects in the presence of
imperfect compliance (e.g., under an irregular assignment mechanism). We show,
through Monte Carlo simulations, that the proposed Bayesian Causal Forest with
Instrumental Variable (BCF-IV) methodology outperforms other machine learning
techniques tailored for causal inference in discovering and estimating the
heterogeneous causal effects. BCF-IV sheds a light on the heterogeneity of
causal effects in instrumental variable scenarios and, in turn, provides the
policy-makers with a relevant tool for targeted policies. Its empirical
application evaluates the effects of additional funding on students'
performances. The results indicate that BCF-IV could be used to enhance the
effectiveness of school funding on students' performance. Code is available at
https://github.com/fbargaglistoffi/BCF-IV.
| stat |
Iterated Feature Screening based on Distance Correlation for
Ultrahigh-Dimensional Censored Data with Covariates Measurement Error | Feature screening is an important method to reduce the dimension and capture
informative variables in ultrahigh-dimensional data analysis. Many methods have
been developed for feature screening. These methods, however, are challenged by
complex features pertinent to the data collection as well as the nature of the
data themselves. Typically, incomplete response caused by right-censoring and
covariates measurement error are often accompanying with survival analysis.
Even though there are many methods have been proposed for censored data, little
work has been available when both incomplete response and measurement error
occur simultaneously. In addition, the conventional feature screening methods
may fail to detect the truly important covariates which are marginally
independent of the response variable due to correlations among covariates. In
this paper, we explore this important problem and propose the valid feature
screening method in the presence of survival data with measurement error. In
addition, we also develop the iteration method to improve the accuracy of
selecting all important covariates. Numerical studies are reported to assess
the performance of the proposed method. Finally, we implement the proposed
method to two different real datasets.
| stat |
Model-based simultaneous inference for multiple subgroups and multiple
endpoints | Various methodological options exist on evaluating differences in both
subgroups and the overall population. Most desirable is the simultaneous study
of multiple endpoints in several populations. We investigate a newer method
using multiple marginal models (mmm) which allows flexible handling of multiple
endpoints, including continuous, binary or time-to-event data. This paper
explores the performance of mmm in contrast to the standard Bonferroni approach
via simulation. Mainly these methods are compared on the basis of their
familywise error rate and power under different scenarios, varying in sample
size and standard deviation. Additionally, it is shown that the method can deal
with overlapping subgroup definitions and different combinations of endpoints
may be assumed. The reanalysis of a clinical example shows a practical
application.
| stat |
Confidence intervals with higher accuracy for short and long memory
linear processes | In this paper an easy to implement method of stochastically weighing short
and long memory linear processes is introduced. The method renders
asymptotically exact size confidence intervals for the population mean which
are significantly more accurate than their classical counterparts for each
fixed sample size $n$. It is illustrated both theoretically and numerically
that the randomization framework of this paper produces randomized (asymptotic)
pivotal quantities, for the mean, which admit central limit theorems with
smaller magnitudes of error as compared to those of their leading classical
counterparts. An Edgeworth expansion result for randomly weighted linear
processes whose innovations do not necessarily satisfy the Cramer condition, is
also established.
| stat |
Scalable GWR: A linear-time algorithm for large-scale geographically
weighted regression with polynomial kernels | Although a number of studies have developed fast geographically weighted
regression (GWR) algorithms for large samples, none of them has achieved
linear-time estimation, which is considered a requisite for big data analysis
in machine learning, geostatistics, and related domains. Against this backdrop,
this study proposes a scalable GWR (ScaGWR) for large datasets. The key
improvement is the calibration of the model through a pre-compression of the
matrices and vectors whose size depends on the sample size, prior to the
leave-one-out cross-validation, which is the heaviest computational step in
conventional GWR. This pre-compression allows us to run the proposed GWR
extension so that its computation time increases linearly with the sample size.
With this improvement, the ScaGWR can be calibrated with one million
observations without parallelization. Moreover, the ScaGWR estimator can be
regarded as an empirical Bayesian estimator that is more stable than the
conventional GWR estimator. We compare the ScaGWR with the conventional GWR in
terms of estimation accuracy and computational efficiency using a Monte Carlo
simulation. Then, we apply these methods to a US income analysis. The code for
ScaGWR is available in the R package scgwr. The code is embedded into C++ code
and implemented in another R package, GWmodel.
| stat |
A projection approach for multiple monotone regression | Shape-constrained inference has wide applicability in bioassay, medicine,
economics, risk assessment, and many other fields. Although there has been a
large amount of work on monotone-constrained univariate curve estimation,
multivariate shape-constrained problems are much more challenging, and fewer
advances have been made in this direction. With a focus on monotone regression
with multiple predictors, this current work proposes a projection approach to
estimate a multiple monotone regression function. An initial unconstrained
estimator -- such as a local polynomial estimator or spline estimator -- is
first obtained, which is then projected onto the shape-constrained space. A
shape-constrained estimate (with multiple predictors) is obtained by
sequentially projecting an (adjusted) initial estimator along each univariate
direction. Compared to the initial unconstrained estimator, the projection
estimate results in a reduction of estimation error in terms of both $L^p$
($p\geq 1$) distance and supremum distance. We also derive the asymptotic
distribution of the projection estimate. Simple computational algorithms are
available for implementing the projection in both the unidimensional and higher
dimensional cases. Our work provides a simple recipe for practitioners to use
in real applications, and is illustrated with a joint-action example from
environmental toxicology.
| stat |
Conformal Inference of Counterfactuals and Individual Treatment Effects | Evaluating treatment effect heterogeneity widely informs treatment decision
making. At the moment, much emphasis is placed on the estimation of the
conditional average treatment effect via flexible machine learning algorithms.
While these methods enjoy some theoretical appeal in terms of consistency and
convergence rates, they generally perform poorly in terms of uncertainty
quantification. This is troubling since assessing risk is crucial for reliable
decision-making in sensitive and uncertain environments. In this work, we
propose a conformal inference-based approach that can produce reliable interval
estimates for counterfactuals and individual treatment effects under the
potential outcome framework. For completely randomized or stratified randomized
experiments with perfect compliance, the intervals have guaranteed average
coverage in finite samples regardless of the unknown data generating mechanism.
For randomized experiments with ignorable compliance and general observational
studies obeying the strong ignorability assumption, the intervals satisfy a
doubly robust property which states the following: the average coverage is
approximately controlled if either the propensity score or the conditional
quantiles of potential outcomes can be estimated accurately. Numerical studies
on both synthetic and real datasets empirically demonstrate that existing
methods suffer from a significant coverage deficit even in simple models. In
contrast, our methods achieve the desired coverage with reasonably short
intervals.
| stat |
Conditional Kernel Density Estimation Considering Autocorrelation for
Renewable Energy Probabilistic Modeling | Renewable energy is essential for energy security and global warming
mitigation. However, power generation from renewable energy sources is
uncertain due to volatile weather conditions and complex equipment operations.
To improve equipment's operation efficiency, it is important to understand and
characterize the uncertainty in renewable power generation. In this paper, we
proposed a conditional kernel density estimation method to model the
distribution of equipment's power output given any weather conditions. It
explicitly accounts for the temporal dependence in the data stream and uses an
iterative procedure to reduce the bias in kernel density estimation. Compared
with existing literature, our approach is especially useful for the purposes of
equipment condition monitoring or short-term renewable energy forecasting,
where the data dependence plays a more significant role. We demonstrate our
method and compare it with alternatives through real applications.
| stat |
Prediction & Model Evaluation for Space-Time Data | Evaluation metrics for prediction error, model selection and model averaging
on space-time data are understudied and poorly understood. The absence of
independent replication makes prediction ambiguous as a concept and renders
evaluation procedures developed for independent data inappropriate for most
space-time prediction problems. Motivated by air pollution data collected
during California wildfires in 2008, this manuscript attempts a formalization
of the true prediction error associated with spatial interpolation. We
investigate a variety of cross-validation (CV) procedures employing both
simulations and case studies to provide insight into the nature of the estimand
targeted by alternative data partition strategies. Consistent with recent best
practice, we find that location-based cross-validation is appropriate for
estimating spatial interpolation error as in our analysis of the California
wildfire data. Interestingly, commonly held notions of bias-variance trade-off
of CV fold size do not trivially apply to dependent data, and we recommend
leave-one-location-out (LOLO) CV as the preferred prediction error metric for
spatial interpolation.
| stat |
Combining biomarker and self-reported dietary intake data: a review of
the state of the art and an exposition of concepts | Classical approaches to assessing dietary intake are associated with
measurement error. In an effort to address inherent measurement error in
dietary self-reported data there is increased interest in the use of dietary
biomarkers as objective measures of intake. Furthermore, there is a growing
consensus of the need to combine dietary biomarker data with self-reported
data.
A review of state of the art techniques employed when combining biomarker and
self-reported data is conducted. Two predominant methods, the calibration
method and the method of triads, emerge as relevant techniques used when
combining biomarker and self-reported data to account for measurement errors in
dietary intake assessment. Both methods crucially assume measurement error
independence. To expose and understand the performance of these methods in a
range of realistic settings, their underpinning statistical concepts are
unified and delineated, and thorough simulation studies conducted.
Results show that violation of the methods' assumptions negatively impacts
resulting inference but that this impact is mitigated when the variation of the
biomarker around the true intake is small. Thus there is much scope for the
further development of biomarkers and models in tandem to achieve the ultimate
goal of accurately assessing dietary intake.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.