title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Functional Singular Spectrum Analysis | In this paper, we introduce a new extension of the Singular Spectrum Analysis
(SSA) called functional SSA to analyze functional time series. The new
methodology is developed by integrating ideas from functional data analysis and
univariate SSA. We explore the advantages of the functional SSA in terms of
simulation results and two real data applications. We compare the proposed
approach with Multivariate SSA (MSSA) and dynamic Functional Principal
Component Analysis (dFPCA). The results suggest that further improvement to
MSSA is possible, and the new method provides an attractive alternative to the
dFPCA approach that is used for analyzing correlated functions. We implement
the proposed technique to an application of remote sensing data and a call
center dataset. We have also developed an efficient and user-friendly R package
and a shiny web application to allow interactive exploration of the results.
| stat |
Maximum conditional entropy Hamiltonian Monte Carlo sampler | The performance of Hamiltonian Monte Carlo (HMC) sampler depends critically
on some algorithm parameters such as the total integration time and the
numerical integration stepsize. The parameter tuning is particularly
challenging when the mass matrix of the HMC sampler is adapted. We propose in
this work a Kolmogorov-Sinai entropy (KSE) based design criterion to optimize
these algorithm parameters, which can avoid some potential issues in the often
used jumping-distance based measures. For near-Gaussian distributions, we are
able to derive the optimal algorithm parameters with respect to the KSE
criterion analytically. As a byproduct the KSE criterion also provides a
theoretical justification for the need to adapt the mass matrix in HMC sampler.
Based on the results, we propose an adaptive HMC algorithm, and we then
demonstrate the performance of the proposed algorithm with numerical examples.
| stat |
General framework for testing Poisson-Voronoi assumption for real
microstructures | Modeling microstructures is an interesting problem not just in Materials
Science but also in Mathematics and Statistics. The most basic model for steel
microstructure is the Poisson-Voronoi diagram. It has mathematically attractive
properties and it has been used in the approximation of single phase steel
microstructures. The aim of this paper is to develop methods that can be used
to test whether a real steel microstructure can be approximated by such a
model. Therefore, a general framework for testing the Poisson-Voronoi
assumption based on images of 2D sections of real metals is set out. Following
two different approaches, according to the use or not of periodic boundary
conditions, three different model tests are proposed. The first two are based
on the coefficient of variation and the cumulative distribution function of the
cells area. The third exploits tools from to Topological Data Analysis, such as
persistence landscapes.
| stat |
SimVAE: Simulator-Assisted Training forInterpretable Generative Models | This paper presents a simulator-assisted training method (SimVAE) for
variational autoencoders (VAE) that leads to a disentangled and interpretable
latent space. Training SimVAE is a two-step process in which first a deep
generator network(decoder) is trained to approximate the simulator. During this
step, the simulator acts as the data source or as a teacher network. Then an
inference network (encoder)is trained to invert the decoder. As such, upon
complete training, the encoder represents an approximately inverted simulator.
By decoupling the training of the encoder and decoder we bypass some of the
difficulties that arise in training generative models such as VAEs and
generative adversarial networks (GANs). We show applications of our approach in
a variety of domains such as circuit design, graphics de-rendering and other
natural science problems that involve inference via simulation.
| stat |
Dynamic Tensor Recommender Systems | Recommender systems have been extensively used by the entertainment industry,
business marketing and the biomedical industry. In addition to its capacity of
providing preference-based recommendations as an unsupervised learning
methodology, it has been also proven useful in sales forecasting, product
introduction and other production related businesses. Since some consumers and
companies need a recommendation or prediction for future budget, labor and
supply chain coordination, dynamic recommender systems for precise forecasting
have become extremely necessary. In this article, we propose a new
recommendation method, namely the dynamic tensor recommender system (DTRS),
which aims particularly at forecasting future recommendation. The proposed
method utilizes a tensor-valued function of time to integrate time and
contextual information, and creates a time-varying coefficient model for
temporal tensor factorization through a polynomial spline approximation. Major
advantages of the proposed method include competitive future recommendation
predictions and effective prediction interval estimations. In theory, we
establish the convergence rate of the proposed tensor factorization and
asymptotic normality of the spline coefficient estimator. The proposed method
is applied to simulations and IRI marketing data. Numerical studies demonstrate
that the proposed method outperforms existing methods in terms of future time
forecasting.
| stat |
Characterizing and Learning Equivalence Classes of Causal DAGs under
Interventions | We consider the problem of learning causal DAGs in the setting where both
observational and interventional data is available. This setting is common in
biology, where gene regulatory networks can be intervened on using chemical
reagents or gene deletions. Hauser and B\"uhlmann (2012) previously
characterized the identifiability of causal DAGs under perfect interventions,
which eliminate dependencies between targeted variables and their direct
causes. In this paper, we extend these identifiability results to general
interventions, which may modify the dependencies between targeted variables and
their causes without eliminating them. We define and characterize the
interventional Markov equivalence class that can be identified from general
(not necessarily perfect) intervention experiments. We also propose the first
provably consistent algorithm for learning DAGs in this setting and evaluate
our algorithm on simulated and biological datasets.
| stat |
Elections and statistics: the case of "United Russia", 2009-2020 | This survey contains statistics on elections in Russia published in different
places and available online. This data is discussed from the viewpoint of
statistical model selection. The current version is updated including the
materials up to July, 2020 voting on constitutional changes, Belarus 2020
elections and papers that appeared in 2020; most of the data are not consistent
with the assumption of fair elections.
| stat |
A deep-learning based Bayesian approach to seismic imaging and
uncertainty quantification | Uncertainty quantification is essential when dealing with ill-conditioned
inverse problems due to the inherent nonuniqueness of the solution. Bayesian
approaches allow us to determine how likely an estimation of the unknown
parameters is via formulating the posterior distribution. Unfortunately, it is
often not possible to formulate a prior distribution that precisely encodes our
prior knowledge about the unknown. Furthermore, adherence to handcrafted priors
may greatly bias the outcome of the Bayesian analysis. To address this issue,
we propose to use the functional form of a randomly initialized convolutional
neural network as an implicit structured prior, which is shown to promote
natural images and excludes images with unnatural noise. In order to
incorporate the model uncertainty into the final estimate, we sample the
posterior distribution using stochastic gradient Langevin dynamics and perform
Bayesian model averaging on the obtained samples. Our synthetic numerical
experiment verifies that deep priors combined with Bayesian model averaging are
able to partially circumvent imaging artifacts and reduce the risk of
overfitting in the presence of extreme noise. Finally, we present pointwise
variance of the estimates as a measure of uncertainty, which coincides with
regions that are more difficult to image.
| stat |
Compressive Statistical Learning with Random Feature Moments | We describe a general framework --compressive statistical learning-- for
resource-efficient large-scale learning: the training collection is compressed
in one pass into a low-dimensional sketch (a vector of random empirical
generalized moments) that captures the information relevant to the considered
learning task. A near-minimizer of the risk is computed from the sketch through
the solution of a nonlinear least squares problem. We investigate sufficient
sketch sizes to control the generalization error of this procedure. The
framework is illustrated on compressive PCA, compressive clustering, and
compressive Gaussian mixture Modeling with fixed known variance. The latter two
are further developed in a companion paper.
| stat |
Accounting for missing actors in interaction network inference from
abundance data | Network inference aims at unraveling the dependency structure relating
jointly observed variables. Graphical models provide a general framework to
distinguish between marginal and conditional dependency. Unobserved variables
(missing actors) may induce apparent conditional dependencies.In the context of
count data, we introduce a mixture of Poisson log-normal distributions with
tree-shaped graphical models, to recover the dependency structure, including
missing actors. We design a variational EM algorithm and assess its performance
on synthetic data. We demonstrate the ability of our approach to recover
environmental drivers on two ecological datasets. The corresponding R package
is available from github.com/Rmomal/nestor.
| stat |
Geometric nested sampling: sampling from distributions defined on
non-trivial geometries | Metropolis Hastings nested sampling evolves a Markov chain, accepting new
points along the chain according to a version of the Metropolis Hastings
acceptance ratio, which has been modified to satisfy the nested sampling
likelihood constraint. The geometric nested sampling algorithm I present here
is based on the Metropolis Hastings method, but treats parameters as though
they represent points on certain geometric objects, namely circles, tori and
spheres. For parameters which represent points on a circle or torus, the trial
distribution is "wrapped" around the domain of the posterior distribution such
that samples cannot be rejected automatically when evaluating the Metropolis
ratio due to being outside the sampling domain. Furthermore, this enhances the
mobility of the sampler. For parameters which represent coordinates on the
surface of a sphere, the algorithm transforms the parameters into a Cartesian
coordinate system before sampling which again makes sure no samples are
automatically rejected, and provides a physically intuitive way of the sampling
the parameter space.
| stat |
Regularized Nonlinear Regression for Simultaneously Selecting and
Estimating Key Model Parameters | In system identification, estimating parameters of a model using limited
observations results in poor identifiability. To cope with this issue, we
propose a new method to simultaneously select and estimate sensitive parameters
as key model parameters and fix the remaining parameters to a set of typical
values. Our method is formulated as a nonlinear least squares estimator with
L1-regularization on the deviation of parameters from a set of typical values.
First, we provide consistency and oracle properties of the proposed estimator
as a theoretical foundation. Second, we provide a novel approach based on
Levenberg-Marquardt optimization to numerically find the solution to the
formulated problem. Third, to show the effectiveness, we present an application
identifying a biomechanical parametric model of a head position tracking task
for 10 human subjects from limited data. In a simulation study, the variances
of estimated parameters are decreased by 96.1% as compared to that of the
estimated parameters without L1-regularization. In an experimental study, our
method improves the model interpretation by reducing the number of parameters
to be estimated while maintaining variance accounted for (VAF) at above 82.5%.
Moreover, the variances of estimated parameters are reduced by 71.1% as
compared to that of the estimated parameters without L1-regularization. Our
method is 54 times faster than the standard simplex-based optimization to solve
the regularized nonlinear regression.
| stat |
Scalable Extreme Deconvolution | The Extreme Deconvolution method fits a probability density to a dataset
where each observation has Gaussian noise added with a known sample-specific
covariance, originally intended for use with astronomical datasets. The
existing fitting method is batch EM, which would not normally be applied to
large datasets such as the Gaia catalog containing noisy observations of a
billion stars. We propose two minibatch variants of extreme deconvolution,
based on an online variation of the EM algorithm, and direct gradient-based
optimisation of the log-likelihood, both of which can run on GPUs. We
demonstrate that these methods provide faster fitting, whilst being able to
scale to much larger models for use with larger datasets.
| stat |
Causal Estimation with Functional Confounders | Causal inference relies on two fundamental assumptions: ignorability and
positivity. We study causal inference when the true confounder value can be
expressed as a function of the observed data; we call this setting estimation
with functional confounders (EFC). In this setting, ignorability is satisfied,
however positivity is violated, and causal inference is impossible in general.
We consider two scenarios where causal effects are estimable. First, we discuss
interventions on a part of the treatment called functional interventions and a
sufficient condition for effect estimation of these interventions called
functional positivity. Second, we develop conditions for nonparametric effect
estimation based on the gradient fields of the functional confounder and the
true outcome function. To estimate effects under these conditions, we develop
Level-set Orthogonal Descent Estimation (LODE). Further, we prove error bounds
on LODE's effect estimates, evaluate our methods on simulated and real data,
and empirically demonstrate the value of EFC.
| stat |
Asymptotics of Ridge Regression in Convolutional Models | Understanding generalization and estimation error of estimators for simple
models such as linear and generalized linear models has attracted a lot of
attention recently. This is in part due to an interesting observation made in
machine learning community that highly over-parameterized neural networks
achieve zero training error, and yet they are able to generalize well over the
test samples. This phenomenon is captured by the so called double descent
curve, where the generalization error starts decreasing again after the
interpolation threshold. A series of recent works tried to explain such
phenomenon for simple models. In this work, we analyze the asymptotics of
estimation error in ridge estimators for convolutional linear models. These
convolutional inverse problems, also known as deconvolution, naturally arise in
different fields such as seismology, imaging, and acoustics among others. Our
results hold for a large class of input distributions that include i.i.d.
features as a special case. We derive exact formulae for estimation error of
ridge estimators that hold in a certain high-dimensional regime. We show the
double descent phenomenon in our experiments for convolutional models and show
that our theoretical results match the experiments.
| stat |
An Analysis of Factors Influencing Metro Station Ridership: Insights
from Taipei Metro | Travel demand analysis at the planning stage is important for metro system
development. In practice, travel demand can be affected by various factors.
This paper focuses on investigating the factors influencing Taipei metro
ridership at station level over varying time periods. Ordinary Least Square
(OLS) multiple regression models with backward stepwise feature selection are
employed to identify the influencing factors, including land use, social
economic, accessibility, network structure information, etc. Network structure
factors are creatively quantified based on complex network theory to accurately
measure the related information. To enhance goodness-of-fit, the dummy variable
distinguishing transportation hub is incorporated in the modeling. The main
findings in this paper are three-fold: First, there is no distinct difference
between influencing factors of boarding and those of alighting; Second,
ridership is significantly associated with the number of nearby shopping malls,
distance to city center, days since opening, nearby bus stations and dummy
variable for transportation hub; Finally, the ridership on weekdays is mainly
affected by commuting activities, while the ridership on weekends is driven by
commercial access.
| stat |
A spatial template independent component analysis model for
subject-level brain network estimation and inference | Independent component analysis is commonly applied to functional magnetic
resonance imaging (fMRI) data to extract independent components (ICs)
representing functional brain networks. While ICA produces reliable group-level
estimates, single-subject ICA often produces noisy results. Template ICA (tICA)
is a hierarchical ICA model using empirical population priors to produce
reliable subject-level IC estimates. However, this and other hierarchical ICA
models assume unrealistically that subject effects are spatially independent.
Here, we propose spatial template ICA (stICA), which incorporates spatial
process priors into tICA. This results in greater estimation efficiency of ICs
and subject effects. Additionally, the joint posterior distribution can be used
to identify engaged areas using an excursions set approach. By leveraging
spatial dependencies and avoiding massive multiple comparisons, stICA has high
power to detect true effects. We derive an efficient expectation-maximization
algorithm to obtain maximum likelihood estimates of the model parameters and
posterior moments of the latent fields. Based on analysis of simulated data and
fMRI data from the Human Connectome Project, we find that stICA produces
estimates that are more accurate and reliable than benchmark approaches, and
identifies larger and more reliable areas of engagement. The algorithm is quite
tractable, achieving convergence within 7 hours in our fMRI analysis.
| stat |
Assurance for clinical trial design with normally distributed outcomes:
eliciting uncertainty about variances | The assurance method is growing in popularity in clinical trial planning. The
method involves eliciting a prior distribution for the treatment effect, and
then calculating the probability that a proposed trial will produce a
`successful' outcome. For normally distributed observations, uncertainty about
the variance of the normal distribution also needs to be accounted for, but
there is little guidance in the literature on how to elicit a distribution for
a variance parameter. We present a simple elicitation method, and illustrate
how the elicited distribution is incorporated within an assurance calculation.
We also consider multi-stage trials, where a decision to proceed with a larger
trial will follow from the outcome of a smaller trial; we illustrate the role
of the elicted distribution in assessing the information provided by a proposed
smaller trial. Free software is available for implementing our methods.
| stat |
A Comparison of Single and Multiple Changepoint Techniques for Time
Series Data | This paper describes and compares several prominent single and multiple
changepoint techniques for time series data. Due to their importance in
inferential matters, changepoint research on correlated data has accelerated
recently. Unfortunately, small perturbations in model assumptions can
drastically alter changepoint conclusions; for example, heavy positive
correlation in a time series can be misattributed to a mean shift should
correlation be ignored. This paper considers both single and multiple
changepoint techniques. The paper begins by examining cumulative sum (CUSUM)
and likelihood ratio tests and their variants for the single changepoint
problem; here, various statistics, boundary cropping scenarios, and scaling
methods (e.g., scaling to an extreme value or Brownian Bridge limit) are
compared. A recently developed test based on summing squared CUSUM statistics
over all times is shown to have realistic Type I errors and superior detection
power. The paper then turns to the multiple changepoint setting. Here,
penalized likelihoods drive the discourse, with AIC, BIC, mBIC, and MDL
penalties being considered. Binary and wild binary segmentation techniques are
also compared. We introduce a new distance metric specifically designed to
compare two multiple changepoint segmentations. Algorithmic and computational
concerns are discussed and simulations are provided to support all conclusions.
In the end, the multiple changepoint setting admits no clear methodological
winner, performance depending on the particular scenario. Nonetheless, some
practical guidance will emerge.
| stat |
Learning Sample-Specific Models with Low-Rank Personalized Regression | Modern applications of machine learning (ML) deal with increasingly
heterogeneous datasets comprised of data collected from overlapping latent
subpopulations. As a result, traditional models trained over large datasets may
fail to recognize highly predictive localized effects in favour of weakly
predictive global patterns. This is a problem because localized effects are
critical to developing individualized policies and treatment plans in
applications ranging from precision medicine to advertising. To address this
challenge, we propose to estimate sample-specific models that tailor inference
and prediction at the individual level. In contrast to classical ML models that
estimate a single, complex model (or only a few complex models), our approach
produces a model personalized to each sample. These sample-specific models can
be studied to understand subgroup dynamics that go beyond coarse-grained class
labels. Crucially, our approach does not assume that relationships between
samples (e.g. a similarity network) are known a priori. Instead, we use
unmodeled covariates to learn a latent distance metric over the samples. We
apply this approach to financial, biomedical, and electoral data as well as
simulated data and show that sample-specific models provide fine-grained
interpretations of complicated phenomena without sacrificing predictive
accuracy compared to state-of-the-art models such as deep neural networks.
| stat |
Wasserstein Fair Classification | We propose an approach to fair classification that enforces independence
between the classifier outputs and sensitive information by minimizing
Wasserstein-1 distances. The approach has desirable theoretical properties and
is robust to specific choices of the threshold used to obtain class predictions
from model outputs. We introduce different methods that enable hiding sensitive
information at test time or have a simple and fast implementation. We show
empirical performance against different fairness baselines on several benchmark
fairness datasets.
| stat |
Wasserstein Autoregressive Models for Density Time Series | Data consisting of time-indexed distributions of cross-sectional or intraday
returns have been extensively studied in finance, and provide one example in
which the data atoms consist of serially dependent probability distributions.
Motivated by such data, we propose an autoregressive model for density time
series by exploiting the tangent space structure on the space of distributions
that is induced by the Wasserstein metric. The densities themselves are not
assumed to have any specific parametric form, leading to flexible forecasting
of future unobserved densities. The main estimation targets in the order-$p$
Wasserstein autoregressive model are Wasserstein autocorrelations and the
vector-valued autoregressive parameter. We propose suitable estimators and
establish their asymptotic normality, which is verified in a simulation study.
The new order-$p$ Wasserstein autoregressive model leads to a prediction
algorithm, which includes a data driven order selection procedure. Its
performance is compared to existing prediction procedures via application to
four financial return data sets, where a variety of metrics are used to
quantify forecasting accuracy. For most metrics, the proposed model outperforms
existing methods in two of the data sets, while the best empirical performance
in the other two data sets is attained by existing methods based on functional
transformations of the densities.
| stat |
Inner-product Kernels are Asymptotically Equivalent to Binary Discrete
Kernels | This article investigates the eigenspectrum of the inner product-type kernel
matrix $\sqrt{p} \mathbf{K}=\{f( \mathbf{x}_i^{\sf T}
\mathbf{x}_j/\sqrt{p})\}_{i,j=1}^n $ under a binary mixture model in the high
dimensional regime where the number of data $n$ and their dimension $p$ are
both large and comparable. Based on recent advances in random matrix theory, we
show that, for a wide range of nonlinear functions $f$, the eigenspectrum
behavior is asymptotically equivalent to that of an (at most) cubic function.
This sheds new light on the understanding of nonlinearity in large dimensional
problems. As a byproduct, we propose a simple function prototype valued in $
(-1,0,1) $ that, while reducing both storage memory and running time, achieves
the same (asymptotic) classification performance as any arbitrary function $f$.
| stat |
An Asymptotically Optimal Multi-Armed Bandit Algorithm and
Hyperparameter Optimization | The evaluation of hyperparameters, neural architectures, or data augmentation
policies becomes a critical model selection problem in advanced deep learning
with a large hyperparameter search space. In this paper, we propose an
efficient and robust bandit-based algorithm called Sub-Sampling (SS) in the
scenario of hyperparameter search evaluation. It evaluates the potential of
hyperparameters by the sub-samples of observations and is theoretically proved
to be optimal under the criterion of cumulative regret. We further combine SS
with Bayesian Optimization and develop a novel hyperparameter optimization
algorithm called BOSS. Empirical studies validate our theoretical arguments of
SS and demonstrate the superior performance of BOSS on a number of
applications, including Neural Architecture Search (NAS), Data Augmentation
(DA), Object Detection (OD), and Reinforcement Learning (RL).
| stat |
Principled estimation of regression discontinuity designs | Regression discontinuity designs are frequently used to estimate the causal
effect of election outcomes and policy interventions. In these contexts,
treatment effects are typically estimated with covariates included to improve
efficiency. While including covariates improves precision asymptotically, in
practice, treatment effects are estimated with a small number of observations,
resulting in considerable fluctuations in treatment effect magnitude and
precision depending upon the covariates chosen. This practice thus incentivizes
researchers to select covariates which maximize treatment effect statistical
significance rather than precision. Here, I propose a principled approach for
estimating RDDs which provides a means of improving precision with covariates
while minimizing adverse incentives. This is accomplished by integrating the
adaptive LASSO, a machine learning method, into RDD estimation using an R
package developed for this purpose, adaptiveRDD. Using simulations, I show that
this method significantly improves treatment effect precision, particularly
when estimating treatment effects with fewer than 200 observations.
| stat |
On the use of cross-validation for the calibration of the tuning
parameter in the adaptive lasso | The adaptive lasso is a popular extension of the lasso, which was shown to
generally enjoy better theoretical performance, at no additional computational
cost in comparison to the lasso. The adaptive lasso relies on a weighted
version of the $L_1$-norm penalty used in the lasso, where weights are
typically derived from an initial estimate of the parameter vector.
Irrespective of the method chosen to obtain this initial estimate, the
performance of the corresponding version of the adaptive lasso critically
depends on the value of the tuning parameter, which controls the magnitude of
the weighted $L_1$-norm in the penalized criterion. In this article, we show
that the standard cross-validation, although very popular in this context, has
a severe defect when applied for the calibration of the tuning parameter in the
adaptive lasso. We further propose a simple cross-validation scheme which
corrects this defect. Empirical results from a simulation study confirms the
superiority of our approach, in terms of both support recovery and prediction
error. Although we focus on the adaptive lasso under linear regression models,
our work likely extends to other regression models, as well as to the adaptive
versions of other penalized approaches, including the group lasso, fused lasso,
and data shared lasso
| stat |
On the Estimation of Population Size from a Dependent Triple Record
System | Population size estimation based on capture-recapture experiment under triple
record system is an interesting problem in various fields including
epidemiology, population studies, etc. In many real life scenarios, there
exists inherent dependency between capture and recapture attempts. We propose a
novel model that successfully incorporates the possible dependency and the
associated parameters possess nice interpretations. We provide estimation
methodology for the population size and the associated model parameters based
on maximum likelihood method. The proposed model is applied to analyze real
data sets from public health and census coverage evaluation study. The
performance of the proposed estimate is evaluated through extensive simulation
study and the results are compared with the existing competitors. The results
exhibit superiority of the proposed model over the existing competitors both in
real data analysis and simulation study.
| stat |
Unified Discrete-Time Factor Stochastic Volatility and Continuous-Time
Ito Models for Combining Inference Based on Low-Frequency and High-Frequency | This paper introduces unified models for high-dimensional factor-based Ito
process, which can accommodate both continuous-time Ito diffusion and
discrete-time stochastic volatility (SV) models by embedding the discrete SV
model in the continuous instantaneous factor volatility process. We call it the
SV-Ito model. Based on the series of daily integrated factor volatility matrix
estimators, we propose quasi-maximum likelihood and least squares estimation
methods. Their asymptotic properties are established. We apply the proposed
method to predict future vast volatility matrix whose asymptotic behaviors are
studied. A simulation study is conducted to check the finite sample performance
of the proposed estimation and prediction method. An empirical analysis is
carried out to demonstrate the advantage of the SV-Ito model in volatility
prediction and portfolio allocation problems.
| stat |
All Models are Wrong, but Many are Useful: Learning a Variable's
Importance by Studying an Entire Class of Prediction Models Simultaneously | Variable importance (VI) tools describe how much covariates contribute to a
prediction model's accuracy. However, important variables for one
well-performing model (for example, a linear model
$f(\mathbf{x})=\mathbf{x}^{T}\beta$ with a fixed coefficient vector $\beta$)
may be unimportant for another model. In this paper, we propose model class
reliance (MCR) as the range of VI values across all well-performing model in a
prespecified class. Thus, MCR gives a more comprehensive description of
importance by accounting for the fact that many prediction models, possibly of
different parametric forms, may fit the data well. In the process of deriving
MCR, we show several informative results for permutation-based VI estimates,
based on the VI measures used in Random Forests. Specifically, we derive
connections between permutation importance estimates for a single prediction
model, U-statistics, conditional variable importance, conditional causal
effects, and linear model coefficients. We then give probabilistic bounds for
MCR, using a novel, generalizable technique. We apply MCR to a public data set
of Broward County criminal records to study the reliance of recidivism
prediction models on sex and race. In this application, MCR can be used to help
inform VI for unknown, proprietary models.
| stat |
Contextual Bandits with Stochastic Experts | We consider the problem of contextual bandits with stochastic experts, which
is a variation of the traditional stochastic contextual bandit with experts
problem. In our problem setting, we assume access to a class of stochastic
experts, where each expert is a conditional distribution over the arms given a
context. We propose upper-confidence bound (UCB) algorithms for this problem,
which employ two different importance sampling based estimators for the mean
reward for each expert. Both these estimators leverage information leakage
among the experts, thus using samples collected under all the experts to
estimate the mean reward of any given expert. This leads to instance dependent
regret bounds of $\mathcal{O}\left(\lambda(\pmb{\mu})\mathcal{M}\log T/\Delta
\right)$, where $\lambda(\pmb{\mu})$ is a term that depends on the mean rewards
of the experts, $\Delta$ is the smallest gap between the mean reward of the
optimal expert and the rest, and $\mathcal{M}$ quantifies the information
leakage among the experts. We show that under some assumptions
$\lambda(\pmb{\mu})$ is typically $\mathcal{O}(\log N)$, where $N$ is the
number of experts. We implement our algorithm with stochastic experts generated
from cost-sensitive classification oracles and show superior empirical
performance on real-world datasets, when compared to other state of the art
contextual bandit algorithms.
| stat |
A Convex Parametrization of a New Class of Universal Kernel Functions | The accuracy and complexity of kernel learning algorithms is determined by
the set of kernels over which it is able to optimize. An ideal set of kernels
should: admit a linear parameterization (tractability); be dense in the set of
all kernels (accuracy); and every member should be universal so that the
hypothesis space is infinite-dimensional (scalability). Currently, there is no
class of kernel that meets all three criteria - e.g. Gaussians are not
tractable or accurate; polynomials are not scalable. We propose a new class
that meet all three criteria - the Tessellated Kernel (TK) class. Specifically,
the TK class: admits a linear parameterization using positive matrices; is
dense in all kernels; and every element in the class is universal. This implies
that the use of TK kernels for learning the kernel can obviate the need for
selecting candidate kernels in algorithms such as SimpleMKL and parameters such
as the bandwidth. Numerical testing on soft margin Support Vector Machine (SVM)
problems show that algorithms using TK kernels outperform other kernel learning
algorithms and neural networks. Furthermore, our results show that when the
ratio of the number of training data to features is high, the improvement of TK
over MKL increases significantly.
| stat |
Simple and Scalable Parallelized Bayesian Optimization | In recent years, leveraging parallel and distributed computational resources
has become essential to solve problems of high computational cost. Bayesian
optimization (BO) has shown attractive results in those expensive-to-evaluate
problems such as hyperparameter optimization of machine learning algorithms.
While many parallel BO methods have been developed to search efficiently
utilizing these computational resources, these methods assumed synchronous
settings or were not scalable. In this paper, we propose a simple and scalable
BO method for asynchronous parallel settings. Experiments are carried out with
a benchmark function and hyperparameter optimization of multi-layer
perceptrons, which demonstrate the promising performance of the proposed
method.
| stat |
Bootstrap Cross-validation Improves Model Selection in Pharmacometrics | Cross-validation assesses the predictive ability of a model, allowing one to
rank models accordingly. Although the nonparametric bootstrap is almost always
used to assess the variability of a parameter, it can be used as the basis for
cross-validation if one keeps track of which items were not selected in a given
bootstrap iteration. The items which were selected constitute the training data
and the omitted items constitute the testing data. This bootstrap
cross-validation (BS-CV) allows model selection to be made on the basis of
predictive ability by comparing the median values of ensembles of summary
statistics of testing data. BS-CV is herein demonstrated using several summary
statistics, including a new one termed the simple metric for prediction quality
(SMPQ), and using the warfarin data included in the Monolix distribution with
13 pharmacokinetics (PK) models and 12 pharmacodynamics (PD) models. Of note
the two best PK models by AIC had the worst predictive ability, underscoring
the danger of using single realizations of a random variable (such as AIC) as
the basis for model selection. Using these data BS-CV was able to discriminate
between similar indirect response models (inhibition of input versus
stimulation of output). This could be useful in situations in which the
mechanism of action is unknown (unlike warfarin).
| stat |
Cultural transmission modes of music sampling traditions remain stable
despite delocalization in the digital age | Music sampling is a common practice among hip-hop and electronic producers
that has played a critical role in the development of particular subgenres.
Artists preferentially sample drum breaks, and previous studies have suggested
that these may be culturally transmitted. With the advent of digital sampling
technologies and social media the modes of cultural transmission may have
shifted, and music communities may have become decoupled from geography. The
aim of the current study was to determine whether drum breaks are culturally
transmitted through musical collaboration networks, and to identify the factors
driving the evolution of these networks. Using network-based diffusion analysis
we found strong evidence for the cultural transmission of drum breaks via
collaboration between artists, and identified several demographic variables
that bias transmission. Additionally, using network evolution methods we found
evidence that the structure of the collaboration network is no longer biased by
geographic proximity after the year 2000, and that gender disparity has relaxed
over the same period. Despite the delocalization of communities by the
internet, collaboration remains a key transmission mode of music sampling
traditions. The results of this study provide valuable insight into how
demographic biases shape cultural transmission in complex networks, and how the
evolution of these networks has shifted in the digital age.
| stat |
Network Differential Connectivity Analysis | Identifying differences in networks has become a canonical problem in many
biological applications. Here, we focus on testing whether two Gaussian
graphical models are the same. Existing methods try to accomplish this goal by
either directly comparing their estimated structures, or testing the null
hypothesis that the partial correlation matrices are equal. However, estimation
approaches do not provide measures of uncertainty, e.g., $p$-values, which are
crucial in drawing scientific conclusions. On the other hand, existing testing
approaches could lead to misleading results in some cases. To address these
shortcomings, we propose a qualitative hypothesis testing framework, which
tests whether the connectivity patterns in the two networks are the same. Our
framework is especially appropriate if the goal is to identify nodes or edges
that are differentially connected. No existing approach could test such
hypotheses and provide corresponding measures of uncertainty, e.g., $p$-values.
We investigate theoretical and numerical properties of our proposal and
illustrate its utility in biological applications. Theoretically, we show that
under appropriate conditions, our proposal correctly controls the type-I error
rate in testing the qualitative hypothesis. Empirically, we demonstrate the
performance of our proposal using simulation datasets and applications in
cancer genetics and brain imaging studies.
| stat |
Evaluating methods for Lasso selective inference in biomedical research
by a comparative simulation study | Variable selection for regression models plays a key role in the analysis of
biomedical data. However, inference after selection is not covered by classical
statistical frequentist theory which assumes a fixed set of covariates in the
model. We review two interpretations of inference after selection: the full
model view, in which the parameters of interest are those of the full model on
all predictors, and then focus on the submodel view, in which the parameters of
interest are those of the selected model only. In the context of L1-penalized
regression we compare proposals for submodel inference (selective inference)
via confidence intervals available to applied researchers via software packages
using a simulation study inspired by real data commonly seen in biomedical
studies. Furthermore, we present an exemplary application of these methods to a
publicly available dataset to discuss their practical usability. Our findings
indicate that the frequentist properties of selective confidence intervals are
generally acceptable, but desired coverage levels are not guaranteed in all
scenarios except for the most conservative methods. The choice of inference
method potentially has a large impact on the resulting interval estimates,
thereby necessitating that the user is acutely aware of the goal of inference
in order to interpret and communicate the results. Currently available software
packages are not yet very user friendly or robust which might affect their use
in practice. In summary, we find submodel inference after selection useful for
experienced statisticians to assess the importance of individual selected
predictors in future applications.
| stat |
DoubleML -- An Object-Oriented Implementation of Double Machine Learning
in R | The R package DoubleML implements the double/debiased machine learning
framework of Chernozhukov et al. (2018). It provides functionalities to
estimate parameters in causal models based on machine learning methods. The
double machine learning framework consist of three key ingredients: Neyman
orthogonality, high-quality machine learning estimation and sample splitting.
Estimation of nuisance components can be performed by various state-of-the-art
machine learning methods that are available in the mlr3 ecosystem. DoubleML
makes it possible to perform inference in a variety of causal models, including
partially linear and interactive regression models and their extensions to
instrumental variable estimation. The object-oriented implementation of
DoubleML enables a high flexibility for the model specification and makes it
easily extendable. This paper serves as an introduction to the double machine
learning framework and the R package DoubleML. In reproducible code examples
with simulated and real data sets, we demonstrate how DoubleML users can
perform valid inference based on machine learning methods.
| stat |
Probabilistic Inference for Structural Health Monitoring: New Modes of
Learning from Data | In data-driven SHM, the signals recorded from systems in operation can be
noisy and incomplete. Data corresponding to each of the operational,
environmental, and damage states are rarely available a priori; furthermore,
labelling to describe the measurements is often unavailable. In consequence,
the algorithms used to implement SHM should be robust and adaptive, while
accommodating for missing information in the training-data -- such that new
information can be included if it becomes available. By reviewing novel
techniques for statistical learning (introduced in previous work), it is argued
that probabilistic algorithms offer a natural solution to the modelling of SHM
data in practice. In three case-studies, probabilistic methods are adapted for
applications to SHM signals -- including semi-supervised learning, active
learning, and multi-task learning.
| stat |
Bounds on the conditional and average treatment effect with unobserved
confounding factors | For observational studies, we study the sensitivity of causal inference when
treatment assignments may depend on unobserved confounding factors. We develop
a loss minimization approach that quantifies bounds on the conditional average
treatment effect (CATE) when unobserved confounder have a bounded effect on the
odds of treatment selection. Our approach is scalable and allows flexible use
of model classes, including nonparametric and black-box machine learning
methods. Using these bounds, we propose a related sensitivity analysis for the
average treatment effect (ATE), and develop a semi-parametric framework that
extends/bounds the augmented inverse propensity weighted (AIPW) estimator for
the ATE beyond the assumption that all confounders are observed. By
constructing a Neyman orthogonal score, our estimator is a regular root-n
estimator so long as the nuisance parameters can be estimated at the
$o_p(n^{-1/4})$ rate. We complement our methodological development with
optimality results showing that our proposed bounds are tight in certain cases.
We demonstrate our method on simulated and real data examples, and show
accurate coverage of our confidence intervals in practical finite sample
regimes.
| stat |
Cross validation in sparse linear regression with piecewise continuous
nonconvex penalties and its acceleration | We investigate the signal reconstruction performance of sparse linear
regression in the presence of noise when piecewise continuous nonconvex
penalties are used. Among such penalties, we focus on the SCAD penalty. The
contributions of this study are three-fold: We first present a theoretical
analysis of a typical reconstruction performance, using the replica method,
under the assumption that each component of the design matrix is given as an
independent and identically distributed (i.i.d.) Gaussian variable. This
clarifies the superiority of the SCAD estimator compared with $\ell_1$ in a
wide parameter range, although the nonconvex nature of the penalty tends to
lead to solution multiplicity in certain regions. This multiplicity is shown to
be connected to replica symmetry breaking in the spin-glass theory. We also
show that the global minimum of the mean square error between the estimator and
the true signal is located in the replica symmetric phase. Second, we develop
an approximate formula efficiently computing the cross-validation error without
actually conducting the cross-validation, which is also applicable to the
non-i.i.d. design matrices. It is shown that this formula is only applicable to
the unique solution region and tends to be unstable in the multiple solution
region. We implement instability detection procedures, which allows the
approximate formula to stand alone and resultantly enables us to draw phase
diagrams for any specific dataset. Third, we propose an annealing procedure,
called nonconvexity annealing, to obtain the solution path efficiently.
Numerical simulations are conducted on simulated datasets to examine these
results to verify the theoretical results consistency and the approximate
formula efficiency. Another numerical experiment on a real-world dataset is
conducted; its results are consistent with those of earlier studies using the
$\ell_0$ formulation.
| stat |
$k$-means as a variational EM approximation of Gaussian mixture models | We show that $k$-means (Lloyd's algorithm) is obtained as a special case when
truncated variational EM approximations are applied to Gaussian Mixture Models
(GMM) with isotropic Gaussians. In contrast to the standard way to relate
$k$-means and GMMs, the provided derivation shows that it is not required to
consider Gaussians with small variances or the limit case of zero variances.
There are a number of consequences that directly follow from our approach: (A)
$k$-means can be shown to increase a free energy associated with truncated
distributions and this free energy can directly be reformulated in terms of the
$k$-means objective; (B) $k$-means generalizations can directly be derived by
considering the 2nd closest, 3rd closest etc. cluster in addition to just the
closest one; and (C) the embedding of $k$-means into a free energy framework
allows for theoretical interpretations of other $k$-means generalizations in
the literature. In general, truncated variational EM provides a natural and
rigorous quantitative link between $k$-means-like clustering and GMM clustering
algorithms which may be very relevant for future theoretical and empirical
studies.
| stat |
An information-geometric approach to feature extraction and moment
reconstruction in dynamical systems | We propose a dimension reduction framework for feature extraction and moment
reconstruction in dynamical systems that operates on spaces of probability
measures induced by observables of the system rather than directly in the
original data space of the observables themselves as in more conventional
methods. Our approach is based on the fact that orbits of a dynamical system
induce probability measures over the measurable space defined by (partial)
observations of the system. We equip the space of these probability measures
with a divergence, i.e., a distance between probability distributions, and use
this divergence to define a kernel integral operator. The eigenfunctions of
this operator create an orthonormal basis of functions that capture different
timescales of the dynamical system. One of our main results shows that the
evolution of the moments of the dynamics-dependent probability measures can be
related to a time-averaging operator on the original dynamical system. Using
this result, we show that the moments can be expanded in the eigenfunction
basis, thus opening up the avenue for nonparametric forecasting of the moments.
If the collection of probability measures is itself a manifold, we can in
addition equip the statistical manifold with the Riemannian metric and use
techniques from information geometry. We present applications to ergodic
dynamical systems on the 2-torus and the Lorenz 63 system, and show on a
real-world example that a small number of eigenvectors is sufficient to
reconstruct the moments (here the first four moments) of an atmospheric time
series, i.e., the realtime multivariate Madden-Julian oscillation index.
| stat |
Towards Patient Record Summarization Through Joint Phenotype Learning in
HIV Patients | Identifying a patient's key problems over time is a common task for providers
at the point care, yet a complex and time-consuming activity given current
electric health records. To enable a problem-oriented summarizer to identify a
patient's comprehensive list of problems and their salience, we propose an
unsupervised phenotyping approach that jointly learns a large number of
phenotypes/problems across structured and unstructured data. To identify the
appropriate granularity of the learned phenotypes, the model is trained on a
target patient population of the same clinic. To enable the content
organization of a problem-oriented summarizer, the model identifies phenotype
relatedness as well. The model leverages a correlated-mixed membership approach
with variational inference applied to heterogenous clinical data. In this
paper, we focus our experiments on assessing the learned phenotypes and their
relatedness as learned from a specific patient population. We ground our
experiments in phenotyping patients from an HIV clinic in a large urban care
institution (n=7,523), where patients have voluminous, longitudinal
documentation, and where providers would benefit from summaries of these
patient's medical histories, whether about their HIV or any comorbidities. We
find that the learned phenotypes and their relatedness are clinically valid
when assessed qualitatively by clinical experts, and that the model surpasses
baseline in inferring phenotype-relatedness when comparing to existing
expert-curated condition groupings.
| stat |
Modelling Extremes of Spatial Aggregates of Precipitation using
Conditional Methods | Inference on the extremal behaviour of spatial aggregates of precipitation is
important for quantifying river flood risk. There are two classes of previous
approach, with one failing to ensure self-consistency in inference across
different regions of aggregation and the other requiring highly inflexible
marginal and spatial dependence structure assumptions. To overcome these
issues, we propose a model for high-resolution precipitation data, from which
we can simulate realistic fields and explore the behaviour of spatial
aggregates. Recent developments in spatial extremes literature have seen
promising progress with spatial extensions of the Heffernan and Tawn (2004)
model for conditional multivariate extremes, which can handle a wide range of
dependence structures. Our contribution is twofold: new parametric forms for
the dependence parameters of this model; and a novel framework for deriving
aggregates addressing edge effects and sub-regions without rain. We apply our
modelling approach to gridded East-Anglia, UK precipitation data. Return-level
curves for spatial aggregates over different regions of various sizes are
estimated and shown to fit very well to the data.
| stat |
Constrained Minimum Energy Designs | Space-filling designs are important in computer experiments, which are
critical for building a cheap surrogate model that adequately approximates an
expensive computer code. Many design construction techniques in the existing
literature are only applicable for rectangular bounded space, but in real world
applications, the input space can often be non-rectangular because of
constraints on the input variables. One solution to generate designs in a
constrained space is to first generate uniformly distributed samples in the
feasible region, and then use them as the candidate set to construct the
designs. Sequentially Constrained Monte Carlo (SCMC) is the state-of-the-art
technique for candidate generation, but it still requires large number of
constraint evaluations, which is problematic especially when the constraints
are expensive to evaluate. Thus, to reduce constraint evaluations and improve
efficiency, we propose the Constrained Minimum Energy Design (CoMinED) that
utilizes recent advances in deterministic sampling methods. Extensive
simulation results on 15 benchmark problems with dimensions ranging from 2 to
13 are provided for demonstrating the improved performance of CoMinED over the
existing methods.
| stat |
Deep Historical Borrowing Framework to Prospectively and Simultaneously
Synthesize Control Information in Confirmatory Clinical Trials with Multiple
Endpoints | In current clinical trial development, historical information is receiving
more attention as providing value beyond sample size calculation.
Meta-analytic-predictive (MAP) priors and robust MAP priors have been proposed
for prospectively borrowing historical data on a single endpoint. To
simultaneously synthesize control information from multiple endpoints in
confirmatory clinical trials, we propose to approximate posterior probabilities
from a Bayesian hierarchical model and estimate critical values by deep
learning to construct pre-specified decision functions before the trial
conduct. Simulation studies and a case study demonstrate that our method
additionally preserves power, and has a satisfactory performance under
prior-data conflict.
| stat |
Controlling the False Discovery Rate for Binary Feature Selection via
Knockoff | Variable selection has been widely used in data analysis for the past
decades, and it becomes increasingly important in the Big Data era as there are
usually hundreds of variables available in a dataset. To enhance
interpretability of a model, identifying potentially relevant features is often
a step before fitting all the features into a regression model. A good variable
selection method should effectively control the fraction of false discoveries
and ensure large enough power of its selection set. In a lot of contemporary
data applications, a great portion of features are coded as binary variables.
Binary features are widespread in many fields, from online controlled
experiments to genome science to physical statistics. Although there has
recently been a handful of literature for provable false discovery rate (FDR)
control in variable selection, most of the theoretical analyses were based on
some strong dependency assumption or Gaussian assumption among features. In
this paper we propose a variable selection method in regression framework for
selecting binary features. Under mild conditions, we show that FDR is
controlled exactly under a target level in a finite sample if the underlying
distribution of the binary features is known. We show in simulations that FDR
control is still attained when feature distribution is estimated from data. We
also provide theoretical results on the power of our variables selection method
in a linear regression model or a logistic regression model. In the restricted
settings where competitors exist, we show in simulations and real data
application on a HIV antiretroviral therapy dataset that our method has higher
power than the competitor.
| stat |
Stochastic Gradient MCMC for Nonlinear State Space Models | State space models (SSMs) provide a flexible framework for modeling complex
time series via a latent stochastic process. Inference for nonlinear,
non-Gaussian SSMs is often tackled with particle methods that do not scale well
to long time series. The challenge is two-fold: not only do computations scale
linearly with time, as in the linear case, but particle filters additionally
suffer from increasing particle degeneracy with longer series. Stochastic
gradient MCMC methods have been developed to scale inference for hidden Markov
models (HMMs) and linear SSMs using buffered stochastic gradient estimates to
account for temporal dependencies. We extend these stochastic gradient
estimators to nonlinear SSMs using particle methods. We present error bounds
that account for both buffering error and particle error in the case of
nonlinear SSMs that are log-concave in the latent process. We evaluate our
proposed particle buffered stochastic gradient using SGMCMC for inference on
both long sequential synthetic and minute-resolution financial returns data,
demonstrating the importance of this class of methods.
| stat |
Generalized Four Moment Theorem and an Application to CLT for Spiked
Eigenvalues of Large-dimensional Covariance Matrices | We consider a more generalized spiked covariance matrix $\Sigma$, which is a
general non-definite matrix with the spiked eigenvalues scattered into a few
bulks and the largest ones allowed to tend to infinity. By relaxing the
matching of the 4th moment to a tail probability decay, a {\it Generalized Four
Moment Theorem} (G4MT) is proposed to show the universality of the asymptotic
law for the local spectral statistics of generalized spiked covariance
matrices, which implies the limiting distribution of the spiked eigenvalues of
the generalized spiked covariance matrix is independent of the actual
distributions of the samples satisfying our relaxed assumptions. Moreover, by
applying it to the Central Limit Theorem (CLT) for the spiked eigenvalues of
the generalized spiked covariance matrix, we also extend the result of Bai and
Yao (2012) to a general form of the population covariance matrix, where the 4th
moment is not necessarily required to exist and the spiked eigenvalues are
allowed to be dependent on the non-spiked ones, thus meeting the actual cases
better.
| stat |
Modelling the Extremes of Seasonal Viruses and Hospital Congestion: The
Example of Flu in a Swiss Hospital | Viruses causing flu or milder coronavirus colds are often referred to as
"seasonal viruses" as they tend to subside in warmer months. In other words,
meteorological conditions tend to impact the activity of viruses, and this
information can be exploited for the operational management of hospitals. In
this study, we use three years of daily data from one of the biggest hospitals
in Switzerland and focus on modelling the extremes of hospital visits from
patients showing flu-like symptoms and the number of positive cases of flu. We
propose employing a discrete Generalized Pareto distribution for the number of
positive and negative cases, and a Generalized Pareto distribution for the odds
of positive cases. Our modelling framework allows for the parameters of these
distributions to be linked to covariate effects, and for outlying observations
to be dealt with via a robust estimation approach. Because meteorological
conditions may vary over time, we use meteorological and not calendar
variations to explain hospital charge extremes, and our empirical findings
highlight their significance. We propose a measure of hospital congestion and a
related tool to estimate the resulting CaRe (Charge-at-Risk-estimation) under
different meteorological conditions. The relevant numerical computations can be
easily carried out using the freely available GJRM R package. The introduced
approach could be applied to several types of seasonal disease data such as
those derived from the new virus SARS-CoV-2 and its COVID-19 disease which is
at the moment wreaking havoc worldwide. The empirical effectiveness of the
proposed method is assessed through a simulation study.
| stat |
Learning from Multiple Corrupted Sources, with Application to Learning
from Label Proportions | We study binary classification in the setting where the learner is presented
with multiple corrupted training samples, with possibly different sample sizes
and degrees of corruption, and introduce an approach based on minimizing a
weighted combination of corruption-corrected empirical risks. We establish a
generalization error bound, and further show that the bound is optimized when
the weights are certain interpretable and intuitive functions of the sample
sizes and degrees of corruptions. We then apply this setting to the problem of
learning with label proportions (LLP), and propose an algorithm that enjoys the
most general statistical performance guarantees known for LLP. Experiments
demonstrate the utility of our theory.
| stat |
Heterogeneous causal effects of neighborhood policing in New York City
with staggered adoption of the policy | Communities often self select into implementing a regulatory policy, and
adopt the policy at different time points. In New York City, neighborhood
policing was adopted at the police precinct level over the years 2015-2018, and
it is of interest to both (1) evaluate the impact of the policy, and (2)
understand what types of communities are most impacted by the policy, raising
questions of heterogeneous treatment effects. We develop novel statistical
approaches that are robust to unmeasured confounding bias to study the causal
effect of policies implemented at the community level. Using techniques from
high-dimensional Bayesian time-series modeling, we estimate treatment effects
by predicting counterfactual values of what would have happened in the absence
of neighborhood policing. We couple the posterior predictive distribution of
the treatment effect with flexible modeling to identify how the impact of the
policy varies across time and community characteristics. Using pre-treatment
data from New York City, we show our approach produces unbiased estimates of
treatment effects with valid measures of uncertainty. Lastly, we find that
neighborhood policing decreases discretionary arrests, but has little effect on
crime or racial disparities in arrest rates.
| stat |
Comment on "Under-reported data analysis with INAR-hidden Markov chains" | In Fernandez-Fontelo et al (Statis. Med. 2016, DOI 10.1002/sim.7026) hidden
integer-valued autoregressive (INAR) processes are used to estimate reporting
probabilities for various diseases. In this comment it is demonstrated that the
Poisson INAR(1) model with time-homogeneous underreporting can be expressed
equivalently as a completely observed INAR(inf) model with a geometric lag
structure. This implies that estimated reporting probabilities depend on the
assumed lag structure of the latent process.
| stat |
Two-sample Testing on Latent Distance Graphs With Unknown Link Functions | We propose a valid and consistent test for the hypothesis that two latent
distance random graphs on the same vertex set have the same generating latent
positions, up to some unidentifiable similarity transformations. Our test
statistic is based on first estimating the edge probabilities matrices by
truncating the singular value decompositions of the averaged adjacency matrices
in each population and then computing a Spearman rank correlation coefficient
between these estimates. Experimental results on simulated data indicate that
the test procedure has power even when there is only one sample from each
population, provided that the number of vertices is not too small. Application
on a dataset of neural connectome graphs showed that we can distinguish between
scans from different age groups while application on a dataset of epileptogenic
recordings showed that we can discriminate between seizure and non-seizure
events.
| stat |
Causal Discovery with a Mixture of DAGs | Causal processes in biomedicine may contain cycles, evolve over time or
differ between populations. However, many graphical models cannot accommodate
these conditions. We propose to model causation using a mixture of directed
cyclic graphs (DAGs), where the joint distribution in a population follows a
DAG at any single point in time but potentially different DAGs across time. We
also introduce an algorithm called Causal Inference over Mixtures that uses
longitudinal data to infer a graph summarizing the causal relations generated
from a mixture of DAGs. Experiments demonstrate improved performance compared
to prior approaches.
| stat |
A flexible sensitivity analysis approach for unmeasured confounding with
multiple treatments and a binary outcome | In the absence of a randomized experiment, a key assumption for drawing
causal inference about treatment effects is the ignorable treatment assignment.
Violations of the ignorability assumption may lead to biased treatment effect
estimates. Sensitivity analysis helps gauge how causal conclusions will be
altered in response to different magnitude of departure from the ignorability
assumption. However, sensitivity analysis approaches for causal inference with
multiple treatments and binary outcomes are scarce. We propose a flexible Monte
Carlo sensitivity analysis approach for the complex multiple treatment settings
with binary outcomes. We first derive the general bias form introduced by
unmeasured confounding (UMC), with emphasis on theoretical properties uniquely
relevant to multiple treatments. We then propose methods to encode the impact
of UMC on potential outcomes and adjust the estimates of causal effects in
which the presumed UMC is removed. Our proposed methods embed nested multiple
imputation within the Bayesian framework, which allow for seamless integration
of the uncertainty about the sensitivity parameters and sampling variability,
as well as use of reputable Bayesian machine learning techniques for modeling
flexibility. Expansive simulations validate our methods and gain insight into
sensitivity analysis with multiple treatments, and we use the SEER-Medicare
data to demonstrate sensitivity analysis using three treatments for early stage
non-small cell lung cancer. The methods developed in this work are readily
available in the R package SAMTx.
| stat |
Simultaneous inference of periods and period-luminosity relations for
Mira variable stars | The Period--Luminosity relation (PLR) of Mira variable stars is an important
tool to determine astronomical distances. The common approach of estimating the
PLR is a two-step procedure that first estimates the Mira periods and then runs
a linear regression of magnitude on log period. When the light curves are
sparse and noisy, the accuracy of period estimation decreases and can suffer
from aliasing effects. Some methods improve accuracy by incorporating complex
model structures at the expense of significant computational costs. Another
drawback of existing methods is that they only provide point estimation without
proper estimation of uncertainty. To overcome these challenges, we develop a
hierarchical Bayesian model that simultaneously models the quasi-periodic
variations for a collection of Mira light curves while estimating their common
PLR. By borrowing strengths through the PLR, our method automatically reduces
the aliasing effect, improves the accuracy of period estimation, and is capable
of characterizing the estimation uncertainty. We develop a scalable stochastic
variational inference algorithm for computation that can effectively deal with
the multimodal posterior of period. The effectiveness of the proposed method is
demonstrated through simulations, and an application to observations of Miras
in the Local Group galaxy M33. Without using ad-hoc period correction tricks,
our method achieves a distance estimate of M33 that is consistent with
published work. Our method also shows superior robustness to downsampling of
the light curves.
| stat |
PAC-Bayesian Transportation Bound | Empirically, the PAC-Bayesian analysis is known to produce tight risk bounds
for practical machine learning algorithms. However, in its naive form, it can
only deal with stochastic predictors while such predictors are rarely used and
deterministic predictors often performs well in practice. To fill this gap, we
develop a new generalization error bound, the PAC-Bayesian transportation
bound, unifying the PAC-Bayesian analysis and the chaining method in view of
the optimal transportation. It is the first PAC-Bayesian bound that relates the
risks of any two predictors according to their distance, and capable of
evaluating the cost of de-randomization of stochastic predictors faced with
continuous loss functions. As an example, we give an upper bound on the
de-randomization cost of spectrally normalized neural networks (NNs) to
evaluate how much randomness contributes to the generalization of NNs.
| stat |
Simulation assisted machine learning | Motivation: In a predictive modeling setting, if sufficient details of the
system behavior are known, one can build and use a simulation for making
predictions. When sufficient system details are not known, one typically turns
to machine learning, which builds a black-box model of the system using a large
dataset of input sample features and outputs. We consider a setting which is
between these two extremes: some details of the system mechanics are known but
not enough for creating simulations that can be used to make high quality
predictions. In this context we propose using approximate simulations to build
a kernel for use in kernelized machine learning methods, such as support vector
machines. The results of multiple simulations (under various uncertainty
scenarios) are used to compute similarity measures between every pair of
samples: sample pairs are given a high similarity score if they behave
similarly under a wide range of simulation parameters. These similarity values,
rather than the original high dimensional feature data, are used to build the
kernel.
Results: We demonstrate and explore the simulation based kernel (SimKern)
concept using four synthetic complex systems--three biologically inspired
models and one network flow optimization model. We show that, when the number
of training samples is small compared to the number of features, the SimKern
approach dominates over no-prior-knowledge methods. This approach should be
applicable in all disciplines where predictive models are sought and
informative yet approximate simulations are available.
Availability: The Python SimKern software, the demonstration models (in
MATLAB, R), and the datasets are available at
https://github.com/davidcraft/SimKern.
| stat |
Inferring food intake from multiple biomarkers using a latent variable
model | Metabolomic based approaches have gained much attention in recent years due
to their promising potential to deliver objective tools for assessment of food
intake. In particular, multiple biomarkers have emerged for single foods.
However, there is a lack of statistical tools available for combining multiple
biomarkers to infer food intake. Furthermore, there is a paucity of approaches
for estimating the uncertainty around biomarker based prediction of intake.
Here, to facilitate inference on the relationship between multiple
metabolomic biomarkers and food intake in an intervention study conducted under
the A-DIET research programme, a latent variable model, multiMarker, is
proposed. The proposed model draws on factor analytic and mixture of experts
models, describing intake as a continuous latent variable whose value gives
raise to the observed biomarker values. We employ a mixture of Gaussian
distributions to flexibly model the latent variable. A Bayesian hierarchical
modelling framework provides flexibility to adapt to different biomarker
distributions and facilitates prediction of the latent intake along with its
associated uncertainty.
Simulation studies are conducted to assess the performance of the proposed
multiMarker framework, prior to its application to the motivating application
of quantifying apple intake.
| stat |
The Hessian Screening Rule | Predictor screening rules, which discard predictors from the design matrix
before fitting a model, have had sizable impacts on the speed with which
$\ell_1$-regularized regression problems, such as the lasso, can be solved.
Current state-of-the-art screening rules, however, have difficulties in dealing
with highly-correlated predictors, often becoming too conservative. In this
paper, we present a new screening rule to deal with this issue: the Hessian
Screening Rule. The rule uses second-order information from the model in order
to provide more accurate screening as well as higher-quality warm starts. In
our experiments on $\ell_1$-regularized least-squares (the lasso) and logistic
regression, we show that the rule outperforms all other alternatives in
simulated experiments with high correlation, as well as in the majority of real
datasets that we study.
| stat |
Quantization Algorithms for Random Fourier Features | The method of random projection (RP) is the standard technique in machine
learning and many other areas, for dimensionality reduction, approximate near
neighbor search, compressed sensing, etc. Basically, RP provides a simple and
effective scheme for approximating pairwise inner products and Euclidean
distances in massive data. Closely related to RP, the method of random Fourier
features (RFF) has also become popular, for approximating the Gaussian kernel.
RFF applies a specific nonlinear transformation on the projected data from
random projections. In practice, using the (nonlinear) Gaussian kernel often
leads to better performance than the linear kernel (inner product), partly due
to the tuning parameter $(\gamma)$ introduced in the Gaussian kernel. Recently,
there has been a surge of interest in studying properties of RFF.
After random projections, quantization is an important step for efficient
data storage, computation, and transmission. Quantization for RP has also been
extensive studied in the literature. In this paper, we focus on developing
quantization algorithms for RFF. The task is in a sense challenging due to the
tuning parameter $\gamma$ in the Gaussian kernel. For example, the quantizer
and the quantized data might be tied to each specific tuning parameter
$\gamma$. Our contribution begins with an interesting discovery, that the
marginal distribution of RFF is actually free of the Gaussian kernel parameter
$\gamma$. This small finding significantly simplifies the design of the
Lloyd-Max (LM) quantization scheme for RFF in that there would be only one LM
quantizer for RFF (regardless of $\gamma$). We also develop a variant named
LM$^2$-RFF quantizer, which in certain cases is more accurate. Experiments
confirm that the proposed quantization schemes perform well.
| stat |
Modified Multidimensional Scaling and High Dimensional Clustering | Multidimensional scaling is an important dimension reduction tool in
statistics and machine learning. Yet few theoretical results characterizing its
statistical performance exist, not to mention any in high dimensions. By
considering a unified framework that includes low, moderate and high
dimensions, we study multidimensional scaling in the setting of clustering
noisy data. Our results suggest that, the classical multidimensional scaling
can be modified to further improve the quality of embedded samples, especially
when the noise level increases. To this end, we propose {\it modified
multidimensional scaling} which applies a nonlinear transformation to the
sample eigenvalues. The nonlinear transformation depends on the dimensionality,
sample size and moment of noise. We show that modified multidimensional scaling
followed by various clustering algorithms can achieve exact recovery, i.e., all
the cluster labels can be recovered correctly with probability tending to one.
Numerical simulations and two real data applications lend strong support to our
proposed methodology.
| stat |
A guide to state-space modeling of ecological time series | State-space models (SSMs) are an important modeling framework for analyzing
ecological time series. These hierarchical models are commonly used to model
population dynamics, animal movement, and capture-recapture data, and are now
increasingly being used to model other ecological processes. SSMs are popular
because they are flexible and they model the natural variation in ecological
processes separately from observation error. Their flexibility allows
ecologists to model continuous, count, binary, and categorical data with linear
or nonlinear processes that evolve in discrete or continuous time. Modeling the
two sources of stochasticity separately allows researchers to differentiate
between biological variation (e.g., in birth processes) and imprecision in the
sampling methodology, and generally provides better estimates of the ecological
quantities of interest than if only one source of stochasticity is directly
modeled. Since the introduction of SSMs, a broad range of fitting procedures
have been proposed. However, the variety and complexity of these procedures can
limit the ability of ecologists to formulate and fit their own SSMs. We provide
the knowledge for ecologists to create SSMs that are robust to common, and
often hidden, estimation problems, and the model selection and validation tools
that can help them assess how well their models fit their data. In this paper,
we present a review of SSMs that will provide a strong foundation to ecologists
interested in learning about SSMs, introduce new tools to veteran SSM users,
and highlight promising research directions for statisticians interested in
ecological applications. The review is accompanied by an in-depth tutorial that
demonstrates how SSMs models can be fitted and validated in R. Together, the
review and tutorial present an introduction to SSMs that will help ecologists
to formulate, fit, and validate their models.
| stat |
Nested Sampling Methods | Nested sampling (NS) computes parameter posterior distributions and makes
Bayesian model comparison computationally feasible. Its strengths are the
unsupervised navigation of complex, potentially multi-modal posteriors until a
well-defined termination point. A systematic literature review of nested
sampling algorithms and variants is presented. We focus on complete algorithms,
including solutions to likelihood-restricted prior sampling. A new formulation
of NS is presented, which casts the parameter space exploration as a search on
a tree. Previously published ways of obtaining robust error estimates and
dynamic variations of the number of live points are presented as special cases
of this formulation.
| stat |
BAGEL: A Bayesian Graphical Model for Inferring Drug Effect
Longitudinally on Depression in People with HIV | Access and adherence to antiretroviral therapy (ART) has transformed the face
of HIV infection from a fatal to a chronic disease. However, ART is also known
for its side effects. Studies have reported that ART is associated with
depressive symptomatology. Large-scale HIV clinical databases with individuals'
longitudinal depression records, ART medications, and clinical characteristics
offer researchers unprecedented opportunities to study the effects of ART drugs
on depression over time. We develop BAGEL, a Bayesian graphical model to
investigate longitudinal effects of ART drugs on a range of depressive symptoms
while adjusting for participants' demographic, behavior, and clinical
characteristics, and taking into account the heterogeneous population through a
Bayesian nonparametric prior. We evaluate BAGEL through simulation studies.
Application to a dataset from the Women's Interagency HIV Study yields
interpretable and clinically useful results. BAGEL not only can improve our
understanding of ART drugs effects on disparate depression symptoms, but also
has clinical utility in guiding informed and effective treatment selection to
facilitate precision medicine in HIV.
| stat |
A fusion learning method to subgroup analysis of Alzheimer's disease | Uncovering the heterogeneity in the disease progression of Alzheimer's is a
key factor to disease understanding and treatment development, so that
interventions can be tailored to target the subgroups that will benefit most
from the treatment, which is an important goal of precision medicine. However,
in practice, one top methodological challenge hindering the heterogeneity
investigation is that the true subgroup membership of each individual is often
unknown. In this article, we aim to identify latent subgroups of individuals
who share a common disorder progress over time, to predict latent subgroup
memberships, and to estimate and infer the heterogeneous trajectories among the
subgroups. To achieve these goals, we apply a concave fusion learning method
proposed in Ma and Huang (2017) and Ma et al. (2019) to conduct subgroup
analysis for longitudinal trajectories of the Alzheimer's disease data. The
heterogeneous trajectories are represented by subject-specific unknown
functions which are approximated by B-splines. The concave fusion method can
simultaneously estimate the spline coefficients and merge them together for the
subjects belonging to the same subgroup to automatically identify subgroups and
recover the heterogeneous trajectories. The resulting estimator of the disease
trajectory of each subgroup is supported by an asymptotic distribution. It
provides a sound theoretical basis for further conducting statistical inference
in subgroup analysis..
| stat |
On doubly robust estimation for logistic partially linear models | Consider a logistic partially linear model, in which the logit of the mean of
a binary response is related to a linear function of some covariates and a
nonparametric function of other covariates. We derive simple, doubly robust
estimators of coefficient for the covariates in the linear component of the
partially linear model. Such estimators remain consistent if either a nuisance
model is correctly specified for the nonparametric component, or another
nuisance model is correctly specified for the means of the covariates of
interest given other covariates and the response at a fixed value. In previous
works, conditional density models are needed for the latter purposes unless a
scalar, binary covariate is handled. We also propose two specific doubly robust
estimators: one is locally-efficient like in our class of doubly robust
estimators and the other is numerically and statistically simpler and can
achieve reasonable efficiency especially when the true coefficients are close
to 0.
| stat |
BoXHED: Boosted eXact Hazard Estimator with Dynamic covariates | The proliferation of medical monitoring devices makes it possible to track
health vitals at high frequency, enabling the development of dynamic health
risk scores that change with the underlying readings. Survival analysis, in
particular hazard estimation, is well-suited to analyzing this stream of data
to predict disease onset as a function of the time-varying vitals. This paper
introduces the software package BoXHED (pronounced 'box-head') for
nonparametrically estimating hazard functions via gradient boosting. BoXHED 1.0
is a novel tree-based implementation of the generic estimator proposed in Lee,
Chen, Ishwaran (2017), which was designed for handling time-dependent
covariates in a fully nonparametric manner. BoXHED is also the first publicly
available software implementation for Lee, Chen, Ishwaran (2017). Applying
BoXHED to cardiovascular disease onset data from the Framingham Heart Study
reveals novel interaction effects among known risk factors, potentially
resolving an open question in clinical literature.
| stat |
Nonparametric Trace Regression in High Dimensions via Sign Series
Representation | Learning of matrix-valued data has recently surged in a range of scientific
and business applications. Trace regression is a widely used method to model
effects of matrix predictors and has shown great success in matrix learning.
However, nearly all existing trace regression solutions rely on two
assumptions: (i) a known functional form of the conditional mean, and (ii) a
global low-rank structure in the entire range of the regression function, both
of which may be violated in practice. In this article, we relax these
assumptions by developing a general framework for nonparametric trace
regression models via structured sign series representations of high
dimensional functions. The new model embraces both linear and nonlinear trace
effects, and enjoys rank invariance to order-preserving transformations of the
response. In the context of matrix completion, our framework leads to a
substantially richer model based on what we coin as the "sign rank" of a
matrix. We show that the sign series can be statistically characterized by
weighted classification tasks. Based on this connection, we propose a learning
reduction approach to learn the regression model via a series of classifiers,
and develop a parallelable computation algorithm to implement sign series
aggregations. We establish the excess risk bounds, estimation error rates, and
sample complexities. Our proposal provides a broad nonparametric paradigm to
many important matrix learning problems, including matrix regression, matrix
completion, multi-task learning, and compressed sensing. We demonstrate the
advantages of our method through simulations and two applications, one on brain
connectivity study and the other on high-rank image completion.
| stat |
Inference for extreme earthquake magnitudes accounting for a
time-varying measurement process | Investment in measuring a process more completely or accurately is only
useful if these improvements can be utilised during modelling and inference. We
consider how improvements to data quality over time can be incorporated when
selecting a modelling threshold and in the subsequent inference of an extreme
value analysis. Motivated by earthquake catalogues, we consider variable data
quality in the form of rounded and incompletely observed data. We develop an
approach to select a time-varying modelling threshold that makes best use of
the available data, accounting for uncertainty in the magnitude model and for
the rounding of observations. We show the benefits of the proposed approach on
simulated data and apply the method to a catalogue of earthquakes induced by
gas extraction in the Netherlands. This more than doubles the usable catalogue
size and greatly increases the precision of high magnitude quantile estimates.
This has important consequences for the design and cost of earthquake defences.
For the first time, we find compelling data-driven evidence against the
applicability of the Gutenberg-Richer law to these earthquakes. Furthermore,
our approach to automated threshold selection appears to have much potential
for generic applications of extreme value methods.
| stat |
A Multivariate Extreme Value Theory Approach to Anomaly Clustering and
Visualization | In a wide variety of situations, anomalies in the behaviour of a complex
system, whose health is monitored through the observation of a random vector X
= (X1,. .. , X d) valued in R d , correspond to the simultaneous occurrence of
extreme values for certain subgroups $\alpha$ $\subset$ {1,. .. , d} of
variables Xj. Under the heavy-tail assumption, which is precisely appropriate
for modeling these phenomena, statistical methods relying on multivariate
extreme value theory have been developed in the past few years for identifying
such events/subgroups. This paper exploits this approach much further by means
of a novel mixture model that permits to describe the distribution of extremal
observations and where the anomaly type $\alpha$ is viewed as a latent
variable. One may then take advantage of the model by assigning to any extreme
point a posterior probability for each anomaly type $\alpha$, defining
implicitly a similarity measure between anomalies. It is explained at length
how the latter permits to cluster extreme observations and obtain an
informative planar representation of anomalies using standard graph-mining
tools. The relevance and usefulness of the clustering and 2-d visual display
thus designed is illustrated on simulated datasets and on real observations as
well, in the aeronautics application domain.
| stat |
Is Basketball a Game of Runs? | Basketball is often referred to as "a game of runs." We investigate the
appropriateness of this claim using data from the full NBA 2016-17 season,
comparing actual longest runs of scoring events to what long run theory
predicts under the assumption that team "momentum" is not present. We provide
several different variations of the analysis. Our results consistently indicate
that the lengths of longest runs in NBA games are no longer than those that
would occur naturally when scoring events are generated by a random process,
rather than one that is influenced by "momentum".
| stat |
Combining multiple observational data sources to estimate causal effects | The era of big data has witnessed an increasing availability of multiple data
sources for statistical analyses. We consider estimation of causal effects
combining big main data with unmeasured confounders and smaller validation data
with supplementary information on these confounders. Under the unconfoundedness
assumption with completely observed confounders, the smaller validation data
allow for constructing consistent estimators for causal effects, but the big
main data can only give error-prone estimators in general. However, by
leveraging the information in the big main data in a principled way, we can
improve the estimation efficiencies yet preserve the consistencies of the
initial estimators based solely on the validation data. Our framework applies
to asymptotically normal estimators, including the commonly-used regression
imputation, weighting, and matching estimators, and does not require a correct
specification of the model relating the unmeasured confounders to the observed
variables. We also propose appropriate bootstrap procedures, which makes our
method straightforward to implement using software routines for existing
estimators.
| stat |
Advanced Stationary and Non-Stationary Kernel Designs for Domain-Aware
Gaussian Processes | Gaussian process regression is a widely-applied method for function
approximation and uncertainty quantification. The technique has gained
popularity recently in the machine learning community due to its robustness and
interpretability. The mathematical methods we discuss in this paper are an
extension of the Gaussian-process framework. We are proposing advanced kernel
designs that only allow for functions with certain desirable characteristics to
be elements of the reproducing kernel Hilbert space (RKHS) that underlies all
kernel methods and serves as the sample space for Gaussian process regression.
These desirable characteristics reflect the underlying physics; two obvious
examples are symmetry and periodicity constraints. In addition, non-stationary
kernel designs can be defined in the same framework to yield flexible
multi-task Gaussian processes. We will show the impact of advanced kernel
designs on Gaussian processes using several synthetic and two scientific data
sets. The results show that including domain knowledge, communicated through
advanced kernel designs, has a significant impact on the accuracy and relevance
of the function approximation.
| stat |
Data-Free Likelihood-Informed Dimension Reduction of Bayesian Inverse
Problems | Identifying a low-dimensional informed parameter subspace offers a viable
path to alleviating the dimensionality challenge in the sampled-based solution
to large-scale Bayesian inverse problems. This paper introduces a novel
gradient-based dimension reduction method in which the informed subspace does
not depend on the data. This permits an online-offline computational strategy
where the expensive low-dimensional structure of the problem is detected in an
offline phase, meaning before observing the data. This strategy is particularly
relevant for multiple inversion problems as the same informed subspace can be
reused. The proposed approach allows controlling the approximation error (in
expectation over the data) of the posterior distribution. We also present
sampling strategies that exploit the informed subspace to draw efficiently
samples from the exact posterior distribution. The method is successfully
illustrated on two numerical examples: a PDE-based inverse problem with a
Gaussian process prior and a tomography problem with Poisson data and a
Besov-$\mathcal{B}^2_{11}$ prior.
| stat |
On Pruning for Score-Based Bayesian Network Structure Learning | Many algorithms for score-based Bayesian network structure learning (BNSL),
in particular exact ones, take as input a collection of potentially optimal
parent sets for each variable in the data. Constructing such collections
naively is computationally intensive since the number of parent sets grows
exponentially with the number of variables. Thus, pruning techniques are not
only desirable but essential. While good pruning rules exist for the Bayesian
Information Criterion (BIC), current results for the Bayesian Dirichlet
equivalent uniform (BDeu) score reduce the search space very modestly,
hampering the use of the (often preferred) BDeu. We derive new non-trivial
theoretical upper bounds for the BDeu score that considerably improve on the
state-of-the-art. Since the new bounds are mathematically proven to be tighter
than previous ones and at little extra computational cost, they are a promising
addition to BNSL methods.
| stat |
A Combined Approach To Detect Key Variables In Thick Data Analytics | In machine learning one of the strategic tasks is the selection of only
significant variables as predictors for the response(s). In this paper an
approach is proposed which consists in the application of permutation tests on
the candidate predictor variables in the aim of identifying only the most
informative ones. Several industrial problems may benefit from such an
approach, and an application in the field of chemical analysis is presented. A
comparison is carried out between the approach proposed and Lasso, that is one
of the most common alternatives for feature selection available in the
literature.
| stat |
Estimation of the number of irregular foreigners in Poland using
non-linear count regression models | Population size estimation requires access to unit-level data in order to
correctly apply capture-recapture methods. Unfortunately, for reasons of
confidentiality access to such data may be limited. To overcome this issue we
apply and extend the hierarchical Poisson-Gamma model proposed by Zhang (2008),
which initially was used to estimate the number of irregular foreigners in
Norway.
The model is an alternative to the current capture-recapture approach as it
does not require linking multiple sources and is solely based on aggregated
administrative data that include (1) the number of apprehended irregular
foreigners, (2) the number of foreigners who faced criminal charges and (3) the
number of foreigners registered in the central population register. The model
explicitly assumes a relationship between the unauthorized and registered
population, which is motivated by the interconnection between these two groups.
This makes the estimation conditionally dependent on the size of regular
population, provides interpretation with analogy to registered population and
makes the estimated parameter more stable over time.
In this paper, we modify the original idea to allow for covariates and
flexible count distributions in order to estimate the number of irregular
foreigners in Poland in 2019. We also propose a parametric bootstrap for
estimating standard errors of estimates. Based on the extended model we
conclude that in as of 31.03.2019 and 30.09.2019 around 15,000 and 20,000
foreigners and were residing in Poland without valid permits. This means that
those apprehended by the Polish Border Guard account for around 15-20% of the
total.
| stat |
Approximate computation of projection depths | Data depth is a concept in multivariate statistics that measures the
centrality of a point in a given data cloud in $\IR^d$. If the depth of a point
can be represented as the minimum of the depths with respect to all
one-dimensional projections of the data, then the depth satisfies the so-called
projection property. Such depths form an important class that includes many of
the depths that have been proposed in literature. For depths that satisfy the
projection property an approximate algorithm can easily be constructed since
taking the minimum of the depths with respect to only a finite number of
one-dimensional projections yields an upper bound for the depth with respect to
the multivariate data. Such an algorithm is particularly useful if no exact
algorithm exists or if the exact algorithm has a high computational complexity,
as is the case with the halfspace depth or the projection depth. To compute
these depths in high dimensions, the use of an approximate algorithm with
better complexity is surely preferable. Instead of focusing on a single method
we provide a comprehensive and fair comparison of several methods, both already
described in the literature and original.
| stat |
MissDeepCausal: Causal Inference from Incomplete Data Using Deep Latent
Variable Models | Inferring causal effects of a treatment, intervention or policy from
observational data is central to many applications. However, state-of-the-art
methods for causal inference seldom consider the possibility that covariates
have missing values, which is ubiquitous in many real-world analyses. Missing
data greatly complicate causal inference procedures as they require an adapted
unconfoundedness hypothesis which can be difficult to justify in practice. We
circumvent this issue by considering latent confounders whose distribution is
learned through variational autoencoders adapted to missing values. They can be
used either as a pre-processing step prior to causal inference but we also
suggest to embed them in a multiple imputation strategy to take into account
the variability due to missing values. Numerical experiments demonstrate the
effectiveness of the proposed methodology especially for non-linear models
compared to competitors.
| stat |
A bivariate logistic regression model based on latent variables | Bivariate observations of binary and ordinal data arise frequently and
require a bivariate modelling approach in cases where one is interested in
aspects of the marginal distributions as separate outcomes along with the
association between the two. We consider methods for constructing such
bivariate models with logistic marginals and propose a model based on the
Ali-Mikhail-Haq bivariate logistic distribution. We motivate the model as an
extension of that based on the Gumbel type 2 distribution as considered by
other authors and as a bivariate extension of the logistic distribution which
preserves certain natural characteristics. Basic properties of the obtained
model are studied and the proposed methods are illustrated through analysis of
two data sets, one describing the trekking habits of Norwegian hikers, the
other stemming from a cognitive experiment of visual recognition and awareness.
| stat |
Experimental Designs for Accelerated Degradation Tests Based on Linear
Mixed Effects Models | Accelerated degradation tests are used to provide accurate estimation of
lifetime properties of highly reliable products within a relatively short
testing time. There data from particular tests at high levels of stress
(e.\,g.\ temperature, voltage, or vibration) are extrapolated, through a
physically meaningful model, to obtain estimates of lifetime quantiles under
normal use conditions. In this work, we consider repeated measures accelerated
degradation tests with multiple stress variables, where the degradation paths
are assumed to follow a linear mixed effects model which is quite common in
settings when repeated measures are made. We derive optimal experimental
designs for minimizing the asymptotic variance for estimating the median
failure time under normal use conditions when the time points for measurements
are either fixed in advance or are also to be optimized.
| stat |
Bipartite Interference and Air Pollution Transport: Estimating Health
Effects of Power Plant Interventions | Evaluating air quality interventions is confronted with the challenge of
interference since interventions at a particular pollution source likely impact
air quality and health at distant locations and air quality and health at any
given location are likely impacted by interventions at many sources. The
structure of interference in this context is dictated by complex atmospheric
processes governing how pollution emitted from a particular source is
transformed and transported across space, and can be cast with a bipartite
structure reflecting the two distinct types of units: 1) interventional units
on which treatments are applied or withheld to change pollution emissions; and
2) outcome units on which outcomes of primary interest are measured. We propose
new estimands for bipartite causal inference with interference that construe
two components of treatment: a "key-associated" (or "individual") treatment and
an "upwind" (or "neighborhood") treatment. Estimation is carried out using a
semi-parametric adjustment approach based on joint propensity scores. A
reduced-complexity atmospheric model is deployed to characterize the structure
of the interference network by modeling the movement of air parcels through
time and space. The new methods are deployed to evaluate the effectiveness of
installing flue-gas desulfurization scrubbers on 472 coal-burning power plants
(the interventional units) in reducing Medicare hospitalizations among
22,603,597 Medicare beneficiaries residing across 23,675 ZIP codes in the
United States (the outcome units).
| stat |
Retrospective analysis of a fatal dose-finding trial | The commonplace description of phase 1 clinical trials in oncology as
"primarily concerned with safety" is belied by their near universal adoption of
dose-escalation practices which are inherently unsafe. In contrast with dose
titration, cohort-wise dose escalation regards patients as exchangeable, an
indefensible assumption in the face of widely appreciated inter-individual
heterogeneity in pharmacokinetics and pharmacodynamics (PKPD). I have
previously advanced this argument in terms of a precautionary coherence
principle that brings the well-known coherence notion of Cheung (2005) into
contact with modern imperatives of patient-centeredness and precision dosing.
Here, however, I explore these matters in some mechanistic detail by analyzing
a trial of the bispecific T cell engager AFM11, in which a fatal toxicity
occurred. To this end, I develop a Bayesian dose-response model for a single
ordinal toxicity. By constructing this model's priors to align with the AFM11
trial as designed and conducted, I demonstrate the incompatibility of that
design with any reasonable expectation of safety. Indeed, the model readily
yields prospective estimates of toxic response probabilities that suggest the
fatality in this trial could have been foreseen as likely.
| stat |
Parametric Programming Approach for More Powerful and General Lasso
Selective Inference | Selective Inference (SI) has been actively studied in the past few years for
conducting inference on the features of linear models that are adaptively
selected by feature selection methods such as Lasso. The basic idea of SI is to
make inference conditional on the selection event. Unfortunately, the main
limitation of the original SI approach for Lasso is that the inference is
conducted not only conditional on the selected features but also on their signs
-- this leads to loss of power because of over-conditioning. Although this
limitation can be circumvented by considering the union of such selection
events for all possible combinations of signs, this is only feasible when the
number of selected features is sufficiently small. To address this
computational bottleneck, we propose a parametric programming-based method that
can conduct SI without conditioning on signs even when we have thousands of
active features. The main idea is to compute the continuum path of Lasso
solutions in the direction of a test statistic, and identify the subset of the
data space corresponding to the feature selection event by following the
solution path. The proposed parametric programming-based method not only avoids
the aforementioned computational bottleneck but also improves the performance
and practicality of SI for Lasso in various respects. We conduct several
experiments to demonstrate the effectiveness and efficiency of our proposed
method.
| stat |
Auto-G-Computation of Causal Effects on a Network | Methods for inferring average causal effects have traditionally relied on two
key assumptions: (i) the intervention received by one unit cannot causally
influence the outcome of another; and (ii) units can be organized into
non-overlapping groups such that outcomes of units in separate groups are
independent. In this paper, we develop new statistical methods for causal
inference based on a single realization of a network of connected units for
which neither assumption (i) nor (ii) holds. The proposed approach allows both
for arbitrary forms of interference, whereby the outcome of a unit may depend
on interventions received by other units with whom a network path through
connected units exists; and long range dependence, whereby outcomes for any two
units likewise connected by a path in the network may be dependent. Under
network versions of consistency and no unobserved confounding, inference is
made tractable by an assumption that the network's outcome, treatment and
covariate vectors are a single realization of a certain chain graph model. This
assumption allows inferences about various network causal effects via the
auto-g-computation algorithm, a network generalization of Robins' well-known
g-computation algorithm previously described for causal inference under
assumptions (i) and (ii).
| stat |
A Practical Introduction to Regression Discontinuity Designs:
Foundations | In this Element and its accompanying Element, Matias D. Cattaneo, Nicolas
Idrobo, and Rocio Titiunik provide an accessible and practical guide for the
analysis and interpretation of Regression Discontinuity (RD) designs that
encourages the use of a common set of practices and facilitates the
accumulation of RD-based empirical evidence. In this Element, the authors
discuss the foundations of the canonical Sharp RD design, which has the
following features: (i) the score is continuously distributed and has only one
dimension, (ii) there is only one cutoff, and (iii) compliance with the
treatment assignment is perfect. In the accompanying Element, the authors
discuss practical and conceptual extensions to the basic RD setup.
| stat |
Discussion of "On nearly assumption-free tests of nominal confidence
interval coverage for causal parameters estimated by machine learning" | We congratulate the authors on their exciting paper, which introduces a novel
idea for assessing the estimation bias in causal estimates. Doubly robust
estimators are now part of the standard set of tools in causal inference, but a
typical analysis stops with an estimate and a confidence interval. The authors
give an approach for a unique type of model-checking that allows the user to
check whether the bias is sufficiently small with respect to the standard
error, which is generally required for confidence intervals to be reliable.
| stat |
Archetypal Analysis for Sparse Nonnegative Matrix Factorization:
Robustness Under Misspecification | We consider the problem of sparse nonnegative matrix factorization (NMF) with
archetypal regularization. The goal is to represent a collection of data points
as nonnegative linear combinations of a few nonnegative sparse factors with
appealing geometric properties, arising from the use of archetypal
regularization. We generalize the notion of robustness studied in Javadi and
Montanari (2019) (without sparsity) to the notions of (a) strong robustness
that implies each estimated archetype is close to the underlying archetypes and
(b) weak robustness that implies there exists at least one recovered archetype
that is close to the underlying archetypes. Our theoretical results on
robustness guarantees hold under minimal assumptions on the underlying data,
and applies to settings where the underlying archetypes need not be sparse. We
propose new algorithms for our optimization problem; and present numerical
experiments on synthetic and real datasets that shed further insights into our
proposed framework and theoretical developments.
| stat |
Bayesian Singular Value Regularization via a Cumulative Shrinkage
Process | This study proposes a novel hierarchical prior for inferring possibly
low-rank matrices measured with noise. We consider three-component matrix
factorization, as in singular value decomposition, and its fully Bayesian
inference. The proposed prior is specified by a scale mixture of exponential
distributions that has spike and slab components. The weights for the
spike/slab parts are inferred using a special prior based on a cumulative
shrinkage process. The proposed prior is designed to increasingly aggressively
push less important, or essentially redundant, singular values toward zero,
leading to more accurate estimates of low-rank matrices. To ensure the
parameter identification, we simulate posterior draws from an approximated
posterior, in which the constraints are slightly relaxed, using a No-U-Turn
sampler. By means of a set of simulation studies, we show that our proposal is
competitive with alternative prior specifications and that it does not incur
significant additional computational burden. We apply the proposed approach to
sectoral industrial production in the United States to analyze the structural
change during the Great Moderation period.
| stat |
A Monte Carlo comparison of categorical tests of independence | The $X^2$ and $G^2$ tests are the most frequently applied tests for testing
the independence of two categorical variables. However, no one, to the best of
our knowledge has compared them, extensively, and ultimately answer the
question of which to use and when. Further, their applicability in cases with
zero frequencies has been debated and (non parametric) permutation tests are
suggested. In this work we perform extensive Monte Carlo simulation studies
attempting to answer both aforementioned points. As expected, in large sample
sized cases ($>1,000$) the $X^2$ and $G^2$ are indistinguishable. In the small
sample sized cases ($\leq 1,000$) though, we provide strong evidence supporting
the use of the $X^2$ test regardless of zero frequencies for the case of
unconditional independence. Also, we suggest the use of the permutation based
$G^2$ test for testing conditional independence, at the cost of being
computationally more expensive. The $G^2$ test exhibited inferior performance
and its use should be limited.
| stat |
Hi-CI: Deep Causal Inference in High Dimensions | We address the problem of counterfactual regression using causal inference
(CI) in observational studies consisting of high dimensional covariates and
high cardinality treatments. Confounding bias, which leads to inaccurate
treatment effect estimation, is attributed to covariates that affect both
treatments and outcome. The presence of high-dimensional co-variates
exacerbates the impact of bias as it is harder to isolate and measure the
impact of these confounders. In the presence of high-cardinality treatment
variables, CI is rendered ill-posed due to the increase in the number of
counterfactual outcomes to be predicted. We propose Hi-CI, a deep neural
network (DNN) based framework for estimating causal effects in the presence of
large number of covariates, and high-cardinal and continuous treatment
variables. The proposed architecture comprises of a decorrelation network and
an outcome prediction network. In the decorrelation network, we learn a data
representation in lower dimensions as compared to the original covariates and
addresses confounding bias alongside. Subsequently, in the outcome prediction
network, we learn an embedding of high-cardinality and continuous treatments,
jointly with the data representation. We demonstrate the efficacy of causal
effect prediction of the proposed Hi-CI network using synthetic and real-world
NEWS datasets.
| stat |
Streaming Adaptive Nonparametric Variational Autoencoder | We develop a data driven approach to perform clustering and end-to-end
feature learning simultaneously for streaming data that can adaptively detect
novel clusters in emerging data. Our approach, Adaptive Nonparametric
Variational Autoencoder (AdapVAE), learns the cluster membership through a
Bayesian Nonparametric (BNP) modeling framework with Deep Neural Networks
(DNNs) for feature learning. We develop a joint online variational inference
algorithm to learn feature representations and clustering assignments
simultaneously via iteratively optimizing the Evidence Lower Bound (ELBO). We
resolve the catastrophic forgetting \citep{kirkpatrick2017overcoming}
challenges with streaming data by adopting generative samples from the trained
AdapVAE using previous data, which avoids the need of storing and reusing past
data. We demonstrate the advantages of our model including adaptive novel
cluster detection without discarding useful information learned from past data,
high quality sample generation and comparable clustering performance as
end-to-end batch mode clustering methods on both image and text corpora
benchmark datasets.
| stat |
A Primer on Causality in Data Science | Many questions in Data Science are fundamentally causal in that our objective
is to learn the effect of some exposure, randomized or not, on an outcome
interest. Even studies that are seemingly non-causal, such as those with the
goal of prediction or prevalence estimation, have causal elements, including
differential censoring or measurement. As a result, we, as Data Scientists,
need to consider the underlying causal mechanisms that gave rise to the data,
rather than simply the pattern or association observed in those data. In this
work, we review the 'Causal Roadmap' of Petersen and van der Laan (2014) to
provide an introduction to some key concepts in causal inference. Similar to
other causal frameworks, the steps of the Roadmap include clearly stating the
scientific question, defining of the causal model, translating the scientific
question into a causal parameter, assessing the assumptions needed to express
the causal parameter as a statistical estimand, implementation of statistical
estimators including parametric and semi-parametric methods, and interpretation
of our findings. We believe that using such a framework in Data Science will
help to ensure that our statistical analyses are guided by the scientific
question driving our research, while avoiding over-interpreting our results. We
focus on the effect of an exposure occurring at a single time point and
highlight the use of targeted maximum likelihood estimation (TMLE) with Super
Learner.
| stat |
Importance Weighted Hierarchical Variational Inference | Variational Inference is a powerful tool in the Bayesian modeling toolkit,
however, its effectiveness is determined by the expressivity of the utilized
variational distributions in terms of their ability to match the true posterior
distribution. In turn, the expressivity of the variational family is largely
limited by the requirement of having a tractable density function. To overcome
this roadblock, we introduce a new family of variational upper bounds on a
marginal log density in the case of hierarchical models (also known as latent
variable models). We then give an upper bound on the Kullback-Leibler
divergence and derive a family of increasingly tighter variational lower bounds
on the otherwise intractable standard evidence lower bound for hierarchical
variational distributions, enabling the use of more expressive approximate
posteriors. We show that previously known methods, such as Hierarchical
Variational Models, Semi-Implicit Variational Inference and Doubly
Semi-Implicit Variational Inference can be seen as special cases of the
proposed approach, and empirically demonstrate superior performance of the
proposed method in a set of experiments.
| stat |
Shifting attention to old age: Detecting mortality deceleration using
focused model selection | The decrease in the increase in death rates at old ages is a phenomenon that
has repeatedly been discussed in demographic research. While mortality
deceleration can be explained in the gamma-Gompertz model as an effect of
selection in heterogeneous populations, this phenomenon can be difficult to
assess statistically because it relates to the tail of the distribution of the
ages at death. By using a focused information criterion (FIC) for model
selection, we can directly target model performance at those advanced ages. The
gamma-Gompertz model is reduced to the competing Gompertz model without
mortality deceleration if the variance parameter lies on the boundary of the
parameter space. We develop a new version of the FIC that is adapted to this
non-standard condition. In a simulation study, the new FIC is shown to
outperform other methods in detecting mortality deceleration. The application
of the FIC to extinct French-Canadian birth cohorts demonstrates that focused
model selection can be used to rebut previous assertions about mortality
deceleration.
| stat |
Local biplots for multi-dimensional scaling, with application to the
microbiome | We present local biplots, a an extension of the classic principal components
biplot to multi-dimensional scaling. Noticing that principal components biplots
have an interpretation as the Jacobian of a map from data space to the
principal subspace, we define local biplots as the Jacobian of the analogous
map for multi-dimensional scaling. In the process, we show a close relationship
between our local biplot axes, generalized Euclidean distances, and generalized
principal components. In simulations and real data we show how local biplots
can shed light on what variables or combinations of variables are important for
the low-dimensional embedding provided by multi-dimensional scaling. They give
particular insight into a class of phylogenetically-informed distances commonly
used in the analysis of microbiome data, showing that different variants of
these distances can be interpreted as implicitly smoothing the data along the
phylogenetic tree and that the extent of this smoothing is variable.
| stat |
Disagreement-Based Combinatorial Pure Exploration: Sample Complexity
Bounds and an Efficient Algorithm | We design new algorithms for the combinatorial pure exploration problem in
the multi-arm bandit framework. In this problem, we are given $K$ distributions
and a collection of subsets $\mathcal{V} \subset 2^{[K]}$ of these
distributions, and we would like to find the subset $v \in \mathcal{V}$ that
has largest mean, while collecting, in a sequential fashion, as few samples
from the distributions as possible. In both the fixed budget and fixed
confidence settings, our algorithms achieve new sample-complexity bounds that
provide polynomial improvements on previous results in some settings. Via an
information-theoretic lower bound, we show that no approach based on uniform
sampling can improve on ours in any regime, yielding the first interactive
algorithms for this problem with this basic property. Computationally, we show
how to efficiently implement our fixed confidence algorithm whenever
$\mathcal{V}$ supports efficient linear optimization. Our results involve
precise concentration-of-measure arguments and a new algorithm for linear
programming with exponentially many constraints.
| stat |
Robust Q-learning | Q-learning is a regression-based approach that is widely used to formalize
the development of an optimal dynamic treatment strategy. Finite dimensional
working models are typically used to estimate certain nuisance parameters, and
misspecification of these working models can result in residual confounding
and/or efficiency loss. We propose a robust Q-learning approach which allows
estimating such nuisance parameters using data-adaptive techniques. We study
the asymptotic behavior of our estimators and provide simulation studies that
highlight the need for and usefulness of the proposed method in practice. We
use the data from the "Extending Treatment Effectiveness of Naltrexone"
multi-stage randomized trial to illustrate our proposed methods.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.