title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Repeated out of Sample Fusion in the Estimation of Small Tail
Probabilities | Often, it is required to estimate the probability that a quantity such as
toxicity level, plutonium, temperature, rainfall, damage, wind speed, wave
size, earthquake magnitude, risk, etc., exceeds an unsafe high threshold. The
probability in question is then very small. To estimate such a probability,
information is needed about large values of the quantity of interest. However,
in many cases, the data only contain values below or even far below the
designated threshold, let alone exceedingly large values. It is shown that by
repeated fusion of the data with externally generated random data, more
information about small tail probabilities is obtained with the aid of certain
new statistical functions. This provides relatively short, yet reliable
interval estimates based on moderately large samples. A comparison of the
approach with a method from extreme values theory (Peaks over Threshold, or
POT), using both artificial and real data, points to the merit of repeated out
of sample fusion.
| stat |
Solving Interpretable Kernel Dimension Reduction | Kernel dimensionality reduction (KDR) algorithms find a low dimensional
representation of the original data by optimizing kernel dependency measures
that are capable of capturing nonlinear relationships. The standard strategy is
to first map the data into a high dimensional feature space using kernels prior
to a projection onto a low dimensional space. While KDR methods can be easily
solved by keeping the most dominant eigenvectors of the kernel matrix, its
features are no longer easy to interpret. Alternatively, Interpretable KDR
(IKDR) is different in that it projects onto a subspace \textit{before} the
kernel feature mapping, therefore, the projection matrix can indicate how the
original features linearly combine to form the new features. Unfortunately, the
IKDR objective requires a non-convex manifold optimization that is difficult to
solve and can no longer be solved by eigendecomposition. Recently, an efficient
iterative spectral (eigendecomposition) method (ISM) has been proposed for this
objective in the context of alternative clustering. However, ISM only provides
theoretical guarantees for the Gaussian kernel. This greatly constrains ISM's
usage since any kernel method using ISM is now limited to a single kernel. This
work extends the theoretical guarantees of ISM to an entire family of kernels,
thereby empowering ISM to solve any kernel method of the same objective. In
identifying this family, we prove that each kernel within the family has a
surrogate $\Phi$ matrix and the optimal projection is formed by its most
dominant eigenvectors. With this extension, we establish how a wide range of
IKDR applications across different learning paradigms can be solved by ISM. To
support reproducible results, the source code is made publicly available on
\url{https://github.com/chieh-neu/ISM_supervised_DR}.
| stat |
Functional Ratings in Sports | In this paper, we present a new model for ranking sports teams. Our model
uses all scoring data from all games to produce a functional rating by the
method of least squares. The functional rating can be interpreted as a teams
average point differential adjusted for strength of schedule. Using two team's
functional ratings we can predict the expected point differential at any time
in the game. We looked at three variations of our model accounting for
home-court advantage in different ways. We use the 2018-2019 NCAA Division 1
men's college basketball season to test the models and determined that
home-court advantage is statistically important but does not differ between
teams.
| stat |
Reducing Parameter Space for Neural Network Training | For neural networks (NNs) with rectified linear unit (ReLU) or binary
activation functions, we show that their training can be accomplished in a
reduced parameter space. Specifically, the weights in each neuron can be
trained on the unit sphere, as opposed to the entire space, and the threshold
can be trained in a bounded interval, as opposed to the real line. We show that
the NNs in the reduced parameter space are mathematically equivalent to the
standard NNs with parameters in the whole space. The reduced parameter space
shall facilitate the optimization procedure for the network training, as the
search space becomes (much) smaller. We demonstrate the improved training
performance using numerical examples.
| stat |
Learning Gaussian Graphical Models by symmetric parallel regression
technique | In this contribution we deal with the problem of learning an undirected graph
which encodes the conditional dependence relationship between variables of a
complex system, given a set of observations of this system. This is a very
central problem of modern data analysis and it comes out every time we want to
investigate a deeper relationship between random variables, which is different
from the classical dependence usually measured by the covariance.
In particular, in this contribution we deal with the case of Gaussian
Graphical Models (GGMs) for which the system of variables has a multivariate
gaussian distribution. We study all the existing techniques for such a problem
and propose a smart implementation of the symmetric parallel regression
technique which turns out to be very competitive for learning sparse GGMs under
high dimensional data regime.
| stat |
Sampling by Divergence Minimization | We introduce a family of Markov Chain Monte Carlo (MCMC) methods designed to
sample from target distributions with irregular geometry using an adaptive
scheme. In cases where targets exhibit non-Gaussian behaviour, we propose that
adaption should be regional in nature as opposed to global. Our algorithms
minimize the information projection side of the Kullback-Leibler (KL)
divergence between the proposal distribution class and the target to encourage
proposals distributed similarly to the regional geometry of the target. Unlike
traditional adaptive MCMC, this procedure rapidly adapts to the geometry of the
current position as it explores the space without the need for a large batch of
samples. We extend this approach to multimodal targets by introducing a heavily
tempered chain to enable faster mixing between regions of interest. The
divergence minimization algorithms are tested on target distributions with
multiple irregularly shaped modes and we provide results demonstrating the
effectiveness of our methods.
| stat |
Time Series Modeling on Dynamic Networks | This paper focuses on modeling the dynamic attributes of a dynamic network
with a fixed number of vertices. These attributes are considered as time series
which dependency structure is influenced by the underlying network. They are
modeled by a multivariate doubly stochastic time series framework, that is we
assume linear processes for which the coefficient matrices are stochastic
processes themselves. We explicitly allow for dependence in the dynamics of the
coefficient matrices as well as between these two stochastic processes. This
framework allows for a separate modeling of the attributes and the underlying
network. In this setting, we define network autoregressive models and discuss
their stationarity conditions. Furthermore, an estimation approach is discussed
in a low- and high-dimensional setting and how this can be applied to
forecasting. The finite sample behavior of the forecast approach is
investigated. This approach is applied to real data whereby the goal is to
forecast the GDP of 33 economies.
| stat |
Meta-learning representations for clustering with infinite Gaussian
mixture models | For better clustering performance, appropriate representations are critical.
Although many neural network-based metric learning methods have been proposed,
they do not directly train neural networks to improve clustering performance.
We propose a meta-learning method that train neural networks for obtaining
representations such that clustering performance improves when the
representations are clustered by the variational Bayesian (VB) inference with
an infinite Gaussian mixture model. The proposed method can cluster unseen
unlabeled data using knowledge meta-learned with labeled data that are
different from the unlabeled data. For the objective function, we propose a
continuous approximation of the adjusted Rand index (ARI), by which we can
evaluate the clustering performance from soft clustering assignments. Since the
approximated ARI and the VB inference procedure are differentiable, we can
backpropagate the objective function through the VB inference procedure to
train the neural networks. With experiments using text and image data sets, we
demonstrate that our proposed method has a higher adjusted Rand index than
existing methods do.
| stat |
New insights for the multivariate square-root lasso | We study the multivariate square-root lasso, a method for fitting the
multivariate response (multi-task) linear regression model with dependent
errors. This estimator minimizes the nuclear norm of the residual matrix plus a
convex penalty. Unlike some existing methods for multivariate response linear
regression, which require explicit estimates of the error covariance matrix or
its inverse, the multivariate square-root lasso criterion implicitly accounts
for error dependence and is convex. To justify the use of this estimator, we
establish error bounds which illustrate that like the univariate square-root
lasso, the multivariate square-root lasso is pivotal with respect to the
unknown error covariance matrix. We propose a new algorithm to compute the
estimator: a variation of the alternating direction method of multipliers
algorithm; and discuss an accelerated first order algorithm which can be
applied in certain cases. In both simulation studies and a genomic data
application, we show that the multivariate square-root lasso can outperform
more computationally intensive methods which estimate both the regression
coefficient matrix and error precision matrix.
| stat |
ViLT: Vision-and-Language Transformer Without Convolution or Region
Supervision | Vision-and-Language Pretraining (VLP) has improved performance on various
joint vision-and-language downstream tasks. Current approaches for VLP heavily
rely on image feature extraction processes, most of which involve region
supervisions (e.g., object detection) and the convolutional architecture (e.g.,
ResNet). Although disregarded in the literature, we find it problematic in
terms of both (1) efficiency/speed, that simply extracting input features
requires much more computation than the actual multimodal interaction steps;
and (2) expressive power, as it is upper bounded to the expressive power of the
visual encoder and its predefined visual vocabulary. In this paper, we present
a minimal VLP model, Vision-and-Language Transformer (ViLT), monolithic in the
sense that processing of visual inputs is drastically simplified to just the
same convolution-free manner that we process textual inputs. We show that ViLT
is up to 60 times faster than previous VLP models, yet with competitive or
better downstream task performance.
| stat |
Machine Learning using Stata/Python | We present two related Stata modules, r_ml_stata and c_ml_stata, for fitting
popular Machine Learning (ML) methods both in regression and classification
settings. Using the recent Stata/Python integration platform (sfi) of Stata 16,
these commands provide hyper-parameters' optimal tuning via K-fold
cross-validation using greed search. More specifically, they make use of the
Python Scikit-learn API to carry out both cross-validation and outcome/label
prediction.
| stat |
Robust variable selection for model-based learning in presence of
adulteration | The problem of identifying the most discriminating features when performing
supervised learning has been extensively investigated. In particular, several
methods for variable selection in model-based classification have been
proposed. Surprisingly, the impact of outliers and wrongly labeled units on the
determination of relevant predictors has received far less attention, with
almost no dedicated methodologies available in the literature. In the present
paper, we introduce two robust variable selection approaches: one that embeds a
robust classifier within a greedy-forward selection procedure and the other
based on the theory of maximum likelihood estimation and irrelevance. The
former recasts the feature identification as a model selection problem, while
the latter regards the relevant subset as a model parameter to be estimated.
The benefits of the proposed methods, in contrast with non-robust solutions,
are assessed via an experiment on synthetic data. An application to a
high-dimensional classification problem of contaminated spectroscopic data
concludes the paper.
| stat |
Stormwater on the Margins: Influence of Race, Gender, and Education on
Willingness to Participate in Stormwater Management | Stormwater has immense impacts on urban flooding and water quality, leaving
the marginalized and the impoverished disproportionately impacted by and
vulnerable to stormwater hazards. However, the environmental health concerns of
socially and economically marginalized individuals are largely underestimated.
Through regression analysis of data from three longitudinal surveys, this
article examines how shared experiences of adverse environmental conditions,
based on individuals' race, gender, and education level, shape one's concern
about and willingness to participate in stormwater management. We found that
people of color, women, and less-educated respondents had a greater willingness
to participate in stormwater management than White, male, and more-educated
respondents, and their concern about local stormwater hazards drove their
willingness to participate. This analysis calls attention to how adverse
environmental conditions may shape the perspectives of those experiencing them
and facilitate a greater willingness to participate.
| stat |
Higher-Order Orthogonal Causal Learning for Treatment Effect | Most existing studies on the double/debiased machine learning method
concentrate on the causal parameter estimation recovering from the first-order
orthogonal score function. In this paper, we will construct the
$k^{\mathrm{th}}$-order orthogonal score function for estimating the average
treatment effect (ATE) and present an algorithm that enables us to obtain the
debiased estimator recovered from the score function. Such a higher-order
orthogonal estimator is more robust to the misspecification of the propensity
score than the first-order one does. Besides, it has the merit of being
applicable with many machine learning methodologies such as Lasso, Random
Forests, Neural Nets, etc. We also undergo comprehensive experiments to test
the power of the estimator we construct from the score function using both the
simulated datasets and the real datasets.
| stat |
Parameter Priors for Directed Acyclic Graphical Models and the
Characterization of Several Probability Distributions | We develop simple methods for constructing parameter priors for model choice
among Directed Acyclic Graphical (DAG) models. In particular, we introduce
several assumptions that permit the construction of parameter priors for a
large number of DAG models from a small set of assessments. We then present a
method for directly computing the marginal likelihood of every DAG model given
a random sample with no missing observations. We apply this methodology to
Gaussian DAG models which consist of a recursive set of linear regression
models. We show that the only parameter prior for complete Gaussian DAG models
that satisfies our assumptions is the normal-Wishart distribution. Our analysis
is based on the following new characterization of the Wishart distribution: let
$W$ be an $n \times n$, $n \ge 3$, positive-definite symmetric matrix of random
variables and $f(W)$ be a pdf of $W$. Then, f$(W)$ is a Wishart distribution if
and only if $W_{11} - W_{12} W_{22}^{-1} W'_{12}$ is independent of
$\{W_{12},W_{22}\}$ for every block partitioning $W_{11},W_{12}, W'_{12},
W_{22}$ of $W$. Similar characterizations of the normal and normal-Wishart
distributions are provided as well.
| stat |
TDEFSI: Theory Guided Deep Learning Based Epidemic Forecasting with
Synthetic Information | Influenza-like illness (ILI) places a heavy social and economic burden on our
society. Traditionally, ILI surveillance data is updated weekly and provided at
a spatially coarse resolution. Producing timely and reliable high-resolution
spatiotemporal forecasts for ILI is crucial for local preparedness and optimal
interventions. We present TDEFSI (Theory Guided Deep Learning Based Epidemic
Forecasting with Synthetic Information), an epidemic forecasting framework that
integrates the strengths of deep neural networks and high-resolution
simulations of epidemic processes over networks. TDEFSI yields accurate
high-resolution spatiotemporal forecasts using low-resolution time series data.
During the training phase, TDEFSI uses high-resolution simulations of epidemics
that explicitly model spatial and social heterogeneity inherent in urban
regions as one component of training data. We train a two-branch recurrent
neural network model to take both within-season and between-season
low-resolution observations as features, and output high-resolution detailed
forecasts. The resulting forecasts are not just driven by observed data but
also capture the intricate social, demographic and geographic attributes of
specific urban regions and mathematical theories of disease propagation over
networks. We focus on forecasting the incidence of ILI and evaluate TDEFSI's
performance using synthetic and real-world testing datasets at the state and
county levels in the USA. The results show that, at the state level, our method
achieves comparable/better performance than several state-of-the-art methods.
At the county level, TDEFSI outperforms the other methods. The proposed method
can be applied to other infectious diseases as well.
| stat |
Importance Sampling and its Optimality for Stochastic Simulation Models | We consider the problem of estimating an expected outcome from a stochastic
simulation model. Our goal is to develop a theoretical framework on importance
sampling for such estimation. By investigating the variance of an importance
sampling estimator, we propose a two-stage procedure that involves a regression
stage and a sampling stage to construct the final estimator. We introduce a
parametric and a nonparametric regression estimator in the first stage and
study how the allocation between the two stages affects the performance of the
final estimator. We analyze the variance reduction rates and derive oracle
properties of both methods. We evaluate the empirical performances of the
methods using two numerical examples and a case study on wind turbine
reliability evaluation.
| stat |
Robustly estimating the marginal likelihood for cognitive models via
importance sampling | Recent advances in Markov chain Monte Carlo (MCMC) extend the scope of
Bayesian inference to models for which the likelihood function is intractable.
Although these developments allow us to estimate model parameters, other basic
problems such as estimating the marginal likelihood, a fundamental tool in
Bayesian model selection, remain challenging. This is an important scientific
limitation because testing psychological hypotheses with hierarchical models
has proven difficult with current model selection methods. We propose an
efficient method for estimating the marginal likelihood for models where the
likelihood is intractable, but can be estimated unbiasedly. It is based on
first running a sampling method such as MCMC to obtain samples for the model
parameters, and then using these samples to construct the proposal density in
an importance sampling (IS) framework with an unbiased estimate of the
likelihood. Our method has several attractive properties: it generates an
unbiased estimate of the marginal likelihood, it is robust to the quality and
target of the sampling method used to form the IS proposals, and it is
computationally cheap to estimate the variance of the marginal likelihood
estimator. We also obtain the convergence properties of the method and provide
guidelines on maximizing computational efficiency. The method is illustrated in
two challenging cases involving hierarchical models: identifying the form of
individual differences in an applied choice scenario, and evaluating the best
parameterization of a cognitive model in a speeded decision making context.
Freely available code to implement the methods is provided. Extensions to
posterior moment estimation and parallelization are also discussed.
| stat |
Collocation-Based Output-Error Method for Aircraft System Identification | The output-error method is a mainstay of aircraft system identification from
flight-test data. It is the method of choice for a wide range of applications,
from the estimation of stability and control derivatives for aerodynamic
database generation to sensor bias estimation in flight-path reconstruction.
However, notable limitations of the output-error method are that it requires ad
hoc modifications for applications to unstable systems and it is an iterative
method which is particularly sensitive to the initial guess. In this paper, we
show how to reformulate the estimation as a collocation problem, an approach
common in other disciplines but seldomly used in flight vehicle system
identification. Both formulations are equivalent in terms of having the same
solution, but the collocation-based can be applied without modifications or
loss of efficiency to unstable systems. Examples with simulated and real-world
flight-test data also show that convergence to the optimum is obtained even
with poor initial guesses.
| stat |
Multi Split Conformal Prediction | Split conformal prediction is a computationally efficient method for
performing distribution-free predictive inference in regression. It involves,
however, a one-time random split of the data, and the result depends on the
particular split. To address this problem, we propose multi split conformal
prediction, a simple method based on Markov's inequality to aggregate single
split conformal prediction intervals across multiple splits.
| stat |
DeePathology: Deep Multi-Task Learning for Inferring Molecular Pathology
from Cancer Transcriptome | Despite great advances, molecular cancer pathology is often limited to the
use of a small number of biomarkers rather than the whole transcriptome, partly
due to computational challenges. Here, we introduce a novel architecture of
Deep Neural Networks (DNNs) that is capable of simultaneous inference of
various properties of biological samples, through multi-task and transfer
learning. It encodes the whole transcription profile into a strikingly
low-dimensional latent vector of size 8, and then recovers mRNA and miRNA
expression profiles, tissue and disease type from this vector. This latent
space is significantly better than the original gene expression profiles for
discriminating samples based on their tissue and disease. We employed this
architecture on mRNA transcription profiles of 10787 clinical samples from 34
classes (one healthy and 33 different types of cancer) from 27 tissues. Our
method significantly outperforms prior works and classical machine learning
approaches in predicting tissue-of-origin, normal or disease state and cancer
type of each sample. For tissues with more than one type of cancer, it reaches
99.4\% accuracy in identifying the correct cancer subtype. We also show this
system is very robust against noise and missing values. Collectively, our
results highlight applications of artificial intelligence in molecular cancer
pathology and oncological research. DeePathology is freely available at
\url{https://github.com/SharifBioinf/DeePathology}.
| stat |
Joint Inference of Multiple Graphs from Matrix Polynomials | Inferring graph structure from observations on the nodes is an important and
popular network science task. Departing from the more common inference of a
single graph and motivated by social and biological networks, we study the
problem of jointly inferring multiple graphs from the observation of signals at
their nodes (graph signals), which are assumed to be stationary in the sought
graphs. From a mathematical point of view, graph stationarity implies that the
mapping between the covariance of the signals and the sparse matrix
representing the underlying graph is given by a matrix polynomial. A prominent
example is that of Markov random fields, where the inverse of the covariance
yields the sparse matrix of interest. From a modeling perspective, stationary
graph signals can be used to model linear network processes evolving on a set
of (not necessarily known) networks. Leveraging that matrix polynomials
commute, a convex optimization method along with sufficient conditions that
guarantee the recovery of the true graphs are provided when perfect covariance
information is available. Particularly important from an empirical viewpoint,
we provide high-probability bounds on the recovery error as a function of the
number of signals observed and other key problem parameters. Numerical
experiments using synthetic and real-world data demonstrate the effectiveness
of the proposed method with perfect covariance information as well as its
robustness in the noisy regime.
| stat |
Clustering of Nonnegative Data and an Application to Matrix Completion | In this paper, we propose a simple algorithm to cluster nonnegative data
lying in disjoint subspaces. We analyze its performance in relation to a
certain measure of correlation between said subspaces. We use our clustering
algorithm to develop a matrix completion algorithm which can outperform
standard matrix completion algorithms on data matrices satisfying certain
natural conditions.
| stat |
Optimal survival trees ensemble | Recent studies have adopted an approach of selecting accurate and diverse
trees based on individual or collective performance within an ensemble for
classification and regression problems. This work follows in the wake of these
investigations and considers the possibility of growing a forest of optimal
survival trees. Initially, a large set of survival trees are grown using the
method of random survival forest. The grown trees are then ranked from smallest
to highest value of their prediction error using out-of-bag observations for
each respective survival tree. The top ranked survival trees are then assessed
for their collective performance as an ensemble. This ensemble is initiated
with the survival tree which stands first in rank, then further trees are
tested one by one by adding them to the ensemble in order of rank. A survival
tree is selected for the resultant ensemble if the performance improves after
an assessment using independent training data. This ensemble is called an
optimal trees ensemble (OSTE). The proposed method is assessed using 17
benchmark datasets and the results are compared with those of random survival
forest, conditional inference forest, bagging and a non tree based method, the
Cox proportional hazard model. In addition to improve predictive performance,
the proposed method reduces the number of survival trees in the ensemble as
compared to the other tree based methods. The method is implemented in an R
package called "OSTE".
| stat |
Neyman-Pearson classification: parametrics and sample size requirement | The Neyman-Pearson (NP) paradigm in binary classification seeks classifiers
that achieve a minimal type II error while enforcing the prioritized type I
error controlled under some user-specified level $\alpha$. This paradigm serves
naturally in applications such as severe disease diagnosis and spam detection,
where people have clear priorities among the two error types. Recently, Tong,
Feng and Li (2018) proposed a nonparametric umbrella algorithm that adapts all
scoring-type classification methods (e.g., logistic regression, support vector
machines, random forest) to respect the given type I error upper bound $\alpha$
with high probability, without specific distributional assumptions on the
features and the responses. Universal the umbrella algorithm is, it demands an
explicit minimum sample size requirement on class $0$, which is often the more
scarce class, such as in rare disease diagnosis applications. In this work, we
employ the parametric linear discriminant analysis (LDA) model and propose a
new parametric thresholding algorithm, which does not need the minimum sample
size requirements on class $0$ observations and thus is suitable for small
sample applications such as rare disease diagnosis. Leveraging both the
existing nonparametric and the newly proposed parametric thresholding rules, we
propose four LDA-based NP classifiers, for both low- and high-dimensional
settings. On the theoretical front, we prove NP oracle inequalities for one
proposed classifier, where the rate for excess type II error benefits from the
explicit parametric model assumption. Furthermore, as NP classifiers involve a
sample splitting step of class $0$ observations, we construct a new adaptive
sample splitting scheme that can be applied universally to NP classifiers, and
this adaptive strategy reduces the type II error of these classifiers.
| stat |
Stochastic Convergence Rates and Applications of Adaptive Quadrature in
Bayesian Inference | We provide the first stochastic convergence rates for adaptive Gauss--Hermite
quadrature applied to normalizing the posterior distribution in Bayesian
models. Our results apply to the uniform relative error in the approximate
posterior density, the coverage probabilities of approximate credible sets, and
approximate moments and quantiles, therefore guaranteeing fast asymptotic
convergence of approximate summary statistics used in practice. We demonstrate
via simulation a simple model that matches our stochastic upper bound for small
sample sizes, and apply the method in two challenging low-dimensional examples.
Further, we demonstrate how adaptive quadrature can be used as a crucial
component of a complex Bayesian inference procedure for high-dimensional
parameters. The method is implemented and made publicly available in the
\texttt{aghq} package for the \texttt{R} language.
| stat |
Verbal Autopsy in Civil Registration and Vital Statistics: The
Symptom-Cause Information Archive | The burden of disease is fundamental to understanding, prioritizing, and
monitoring public health interventions. Cause of death is required to calculate
the burden of disease, but in many parts of the developing world deaths are
neither detected nor given a cause. Verbal autopsy is a feasible way of
assigning a cause to a death based on an interview with those who cared for the
person who died. As civil registration and vital statistics systems improve in
the developing world, verbal autopsy is playing and increasingly important role
in providing information about cause of death and the burden of disease. This
note motivates the creation of a global symptom-cause archive containing
reference deaths with both a verbal autopsy and a cause assigned through an
independent mechanism. This archive could provide training and validation data
for refining, developing, and testing machine-based algorithms to automate
cause assignment from verbal autopsy data. This, in turn, would improve the
comparability of machine-assigned causes and provide a means to fine-tune
individual cause assignment within specific contexts.
| stat |
Knockoff Boosted Tree for Model-Free Variable Selection | In this article, we propose a novel strategy for conducting variable
selection without prior model topology knowledge using the knockoff method with
boosted tree models. Our method is inspired by the original knockoff method,
where the differences between original and knockoff variables are used for
variable selection with false discovery rate control. The original method uses
Lasso for regression models and assumes there are more samples than variables.
We extend this method to both model-free and high-dimensional variable
selection. We propose two new sampling methods for generating knockoffs, namely
the sparse covariance and principal component knockoff methods. We test these
methods and compare them with the original knockoff method in terms of their
ability to control type I errors and power. The boosted tree model is a complex
system and has more hyperparameters than models with simpler assumptions. In
our framework, these hyperparameters are either tuned through Bayesian
optimization or fixed at multiple levels for trend detection. In simulation
tests, we also compare the properties and performance of importance test
statistics of tree models. The results include combinations of different
knockoffs and importance test statistics. We also consider scenarios that
include main-effect, interaction, exponential, and second-order models while
assuming the true model structures are unknown. We apply our algorithm for
tumor purity estimation and tumor classification using the Cancer Genome Atlas
(TCGA) gene expression data. The proposed algorithm is included in the KOBT
package, available at
\url{https://cran.r-project.org/web/packages/KOBT/index.html}.
| stat |
PAC-Bayesian theory for stochastic LTI systems | In this paper we derive a PAC-Bayesian error bound for autonomous stochastic
LTI state-space models. The motivation for deriving such error bounds is that
they will allow deriving similar error bounds for more general dynamical
systems, including recurrent neural networks. In turn, PACBayesian error bounds
are known to be useful for analyzing machine learning algorithms and for
deriving new ones.
| stat |
Determining feature importance for actionable climate change mitigation
policies | Given the importance of public support for policy change and implementation,
public policymakers and researchers have attempted to understand the factors
associated with this support for climate change mitigation policy. In this
article, we compare the feasibility of using different supervised learning
methods for regression using a novel socio-economic data set which measures
public support for potential climate change mitigation policies. Following this
model selection, we utilize gradient boosting regression, a well-known
technique in the machine learning community, but relatively uncommon in public
policy and public opinion research, and seek to understand what factors among
the several examined in previous studies are most central to shaping public
support for mitigation policies in climate change studies. The use of this
method provides novel insights into the most important factors for public
support for climate change mitigation policies. Using national survey data, we
find that the perceived risks associated with climate change are more decisive
for shaping public support for policy options promoting renewable energy and
regulating pollutants. However, we observe a very different behavior related to
public support for increasing the use of nuclear energy where climate change
risk perception is no longer the sole decisive feature. Our findings indicate
that public support for renewable energy is inherently different from that for
nuclear energy reliance with the risk perception of climate change, dominant
for the former, playing a subdued role for the latter.
| stat |
Deep Generative Quantile-Copula Models for Probabilistic Forecasting | We introduce a new category of multivariate conditional generative models and
demonstrate its performance and versatility in probabilistic time series
forecasting and simulation. Specifically, the output of quantile regression
networks is expanded from a set of fixed quantiles to the whole Quantile
Function by a univariate mapping from a latent uniform distribution to the
target distribution. Then the multivariate case is solved by learning such
quantile functions for each dimension's marginal distribution, followed by
estimating a conditional Copula to associate these latent uniform random
variables. The quantile functions and copula, together defining the joint
predictive distribution, can be parameterized by a single implicit generative
Deep Neural Network.
| stat |
Machine Learning on EPEX Order Books: Insights and Forecasts | This paper employs machine learning algorithms to forecast German electricity
spot market prices. The forecasts utilize in particular bid and ask order book
data from the spot market but also fundamental market data like renewable
infeed and expected demand. Appropriate feature extraction for the order book
data is developed. Using cross-validation to optimise hyperparameters, neural
networks and random forests are proposed and compared to statistical reference
models. The machine learning models outperform traditional approaches.
| stat |
Tree Tensor Networks for Generative Modeling | Matrix product states (MPS), a tensor network designed for one-dimensional
quantum systems, has been recently proposed for generative modeling of natural
data (such as images) in terms of `Born machine'. However, the exponential
decay of correlation in MPS restricts its representation power heavily for
modeling complex data such as natural images. In this work, we push forward the
effort of applying tensor networks to machine learning by employing the Tree
Tensor Network (TTN) which exhibits balanced performance in expressibility and
efficient training and sampling. We design the tree tensor network to utilize
the 2-dimensional prior of the natural images and develop sweeping learning and
sampling algorithms which can be efficiently implemented utilizing Graphical
Processing Units (GPU). We apply our model to random binary patterns and the
binary MNIST datasets of handwritten digits. We show that TTN is superior to
MPS for generative modeling in keeping correlation of pixels in natural images,
as well as giving better log-likelihood scores in standard datasets of
handwritten digits. We also compare its performance with state-of-the-art
generative models such as the Variational AutoEncoders, Restricted Boltzmann
machines, and PixelCNN. Finally, we discuss the future development of Tensor
Network States in machine learning problems.
| stat |
Multiscale stick-breaking mixture models | We introduce a family of multiscale stick-breaking mixture models for
Bayesian nonparametric density estimation. The Bayesian nonparametric
literature is dominated by single scale methods, exception made for P\`olya
trees and allied approaches. Our proposal is based on a mixture specification
exploiting an infinitely-deep binary tree of random weights that grows
according to a multiscale generalization of a large class of stick-breaking
processes; this multiscale stick-breaking is paired with specific stochastic
processes generating sequences of parameters that induce stochastically ordered
kernel functions. Properties of this family of multiscale stick-breaking
mixtures are described. Focusing on a Gaussian specification, a Markov Chain
Montecarlo algorithm for posterior computation is introduced. The performance
of the method is illustrated analyzing both synthetic and real data sets. The
method is well-suited for data living in $\mathbb{R}$ and is able to detect
densities with varying degree of smoothness and local features.
| stat |
Bayesian Quantile Factor Models | Factor analysis is a flexible technique for assessment of multivariate
dependence and codependence. Besides being an exploratory tool used to reduce
the dimensionality of multivariate data, it allows estimation of common factors
that often have an interesting theoretical interpretation in real problems.
However, in some specific cases the interest involves the effects of latent
factors not only in the mean, but in the entire response distribution,
represented by a quantile. This paper introduces a new class of models, named
quantile factor models, which combines factor model theory with
distribution-free quantile regression producing a robust statistical method.
Bayesian estimation for the proposed model is performed using an efficient
Markov chain Monte Carlo algorithm. The proposed model is evaluated using
synthetic datasets in different settings, in order to evaluate its robustness
and performance under different quantiles compared to more usual methods. The
model is also applied to a financial sector dataset and a heart disease
experiment.
| stat |
Uncertainty and Sensitivity Analyses Methods for Agent-Based
Mathematical Models: An Introductory Review | Multiscale, agent-based mathematical models of biological systems are often
associated with model uncertainty and sensitivity to parameter perturbations.
Here, three uncertainty and sensitivity analyses methods, that are suitable to
use when working with agent-based models, are discussed. These methods are
namely Consistency Analysis, Robustness Analysis and Latin Hypercube Analysis.
This introductory review discusses origins, conventions, implementation and
result interpretation of the aforementioned methods. Information on how to
implement the discussed methods in MATLAB is included.
| stat |
Multi-resolution filters for massive spatio-temporal data | Spatio-temporal data sets are rapidly growing in size. For example,
environmental variables are measured with ever-higher resolution by increasing
numbers of automated sensors mounted on satellites and aircraft. Using such
data, which are typically noisy and incomplete, the goal is to obtain complete
maps of the spatio-temporal process, together with proper uncertainty
quantification. We focus here on real-time filtering inference in linear
Gaussian state-space models. At each time point, the state is a spatial field
evaluated on a very large spatial grid, making exact inference using the Kalman
filter computationally infeasible. Instead, we propose a multi-resolution
filter (MRF), a highly scalable and fully probabilistic filtering method that
resolves spatial features at all scales. We prove that the MRF matrices exhibit
a particular block-sparse multi-resolution structure that is preserved under
filtering operations through time. We also discuss inference on time-varying
parameters using an approximate Rao-Blackwellized particle filter, in which the
integrated likelihood is computed using the MRF. We compare the MRF to existing
approaches in a simulation study and a real satellite-data application.
| stat |
Two-Stage Regularization of Pseudo-Likelihood Estimators with
Application to Time Series | Estimators derived from score functions that are not the likelihood are in
wide use in practical and modern applications. Their regularization is often
carried by pseudo-posterior estimation, equivalently by adding penalty to the
score function. We argue that this approach is suboptimal, and propose a
two-staged alternative involving estimation of a new score function which
better approximates the true likelihood for the purpose of regularization. Our
approach typically identifies with maximum a-posteriori estimation if the
original score function is in fact the likelihood. We apply our theory to
fitting ordinary least squares (OLS) under contemporaneous exogeneity, a
setting appearing often in time series and in which OLS is the estimator of
choice by practitioners.
| stat |
Simultaneous Confidence Band for Stationary Covariance Function of Dense
Functional Data | Inference via simultaneous confidence band is studied for stationary
covariance function of dense functional data. A two-stage estimation procedure
is proposed based on spline approximation, the first stage involving estimation
of all the individual trajectories and the second stage involving estimation of
the covariance function through smoothing the empirical covariance function.
The proposed covariance estimator is smooth and as efficient as the oracle
estimator when all individual trajectories are known. An asymptotic
simultaneous confidence band (SCB) is developed for the true covariance
function, and the coverage probabilities are shown to be asymptotically
correct. Simulation experiments are conducted on the numerical performance of
the proposed estimator and SCB. The proposed method is also illustrated by two
real data examples.
| stat |
Pivotal tests for relevant differences in the second order dynamics of
functional time series | Motivated by the need to statistically quantify differences between modern
(complex) data-sets which commonly result as high-resolution measurements of
stochastic processes varying over a continuum, we propose novel testing
procedures to detect relevant differences between the second order dynamics of
two functional time series. In order to take the between-function dynamics into
account that characterize this type of functional data, a frequency domain
approach is taken. Test statistics are developed to compare differences in the
spectral density operators and in the primary modes of variation as encoded in
the associated eigenelements. Under mild moment conditions, we show convergence
of the underlying statistics to Brownian motions and construct pivotal test
statistics. The latter is essential because the nuisance parameters can be
unwieldy and their robust estimation infeasible, especially if the two
functional time series are dependent. Besides from these novel features, the
properties of the tests are robust to any choice of frequency band enabling
also to compare energy contents at a single frequency. The finite sample
performance of the tests are verified through a simulation study and are
illustrated with an application to fMRI data.
| stat |
On choking and ingestion hazards for children in the United States | The risks of Unintentional Suffocation injuries in the U.S. must be
reconsidered in view of the existent mortality and morbidity statistics. In
particular, fatal injuries due to the Sudden Unexpected Infant Death Syndrome
[SUIDS] should be treated on their own and separated from this group. Because
of a non-appropriate nomenclature, the risks of injuries due to Aspiration
and/or Ingestion of Foreign Bodies have been overestimated in the recent
decades, and should be reconsidered by a factual scrutiny of the statistical
data.
| stat |
Analytic Study of Double Descent in Binary Classification: The Impact of
Loss | Extensive empirical evidence reveals that, for a wide range of different
learning methods and datasets, the risk curve exhibits a double-descent (DD)
trend as a function of the model size. In a recent paper
[Zeyu,Kammoun,Thrampoulidis,2019] the authors studied binary linear
classification models and showed that the test error of gradient descent (GD)
with logistic loss undergoes a DD. In this paper, we complement these results
by extending them to GD with square loss. We show that the DD phenomenon
persists, but we also identify several differences compared to logistic loss.
This emphasizes that crucial features of DD curves (such as their transition
threshold and global minima) depend both on the training data and on the
learning algorithm. We further study the dependence of DD curves on the size of
the training set. Similar to our earlier work, our results are analytic: we
plot the DD curves by first deriving sharp asymptotics for the test error under
Gaussian features. Albeit simple, the models permit a principled study of DD
features, the outcomes of which theoretically corroborate related empirical
findings occurring in more complex learning tasks.
| stat |
Discussion of 'Estimating time-varying causal excursion effect in mobile
health with binary outcomes' by T. Qian et al | We discuss the recent paper on "excursion effect" by T. Qian et al. (2020).
We show that the methods presented have close relationships to others in the
literature, in particular to a series of papers by Robins, Hern\'{a}n and
collaborators on analyzing observational studies as a series of randomized
trials. There is also a close relationship to the history-restricted and the
history-adjusted marginal structural models (MSM). Important differences and
their methodological implications are clarified. We also demonstrate that the
excursion effect can depend on the design and discuss its suitability for
modifying the treatment protocol.
| stat |
The Renyi Gaussian Process: Towards Improved Generalization | We introduce an alternative closed form lower bound on the Gaussian process
($\mathcal{GP}$) likelihood based on the R\'enyi $\alpha$-divergence. This new
lower bound can be viewed as a convex combination of the Nystr\"om
approximation and the exact $\mathcal{GP}$. The key advantage of this bound, is
its capability to control and tune the enforced regularization on the model and
thus is a generalization of the traditional variational $\mathcal{GP}$
regression. From a theoretical perspective, we provide the convergence rate and
risk bound for inference using our proposed approach. Experiments on real data
show that the proposed algorithm may be able to deliver improvement over
several $\mathcal{GP}$ inference methods.
| stat |
Uniformly valid confidence intervals for conditional treatment effects
in misspecified high-dimensional models | Eliminating the effect of confounding in observational studies typically
involves fitting a model for an outcome adjusted for covariates. When, as
often, these covariates are high-dimensional, this necessitates the use of
sparse estimators such as the Lasso, or other regularisation approaches. Naiive
use of such estimators yields confidence intervals for the conditional
treatment effect parameter that are not uniformly valid. Moreover, as the
number of covariates grows with sample size, correctly specifying a model for
the outcome is non-trivial. In this work, we deal with both of these concerns
simultaneously, delivering confidence intervals for conditional treatment
effects that are uniformly valid, regardless of whether the outcome model is
correct. This is done by incorporating an additional model for the
treatment-selection mechanism. When both models are correctly specified, we can
weaken the standard conditions on model sparsity. Our procedure extends to
multivariate treatment effect parameters and complex longitudinal settings.
| stat |
Sequential Importance Sampling With Corrections For Partially Observed
States | We consider an evolving system for which a sequence of observations is being
made, with each observation revealing additional information about current and
past states of the system. We suppose each observation is made without error,
but does not fully determine the state of the system at the time it is made.
Our motivating example is drawn from invasive species biology, where it is
common to know the precise location of invasive organisms that have been
detected by a surveillance program, but at any time during the program there
are invaders that have not been detected.
We propose a sequential importance sampling strategy to infer the state of
the invasion under a Bayesian model of such a system. The strategy involves
simulating multiple alternative states consistent with current knowledge of the
system, as revealed by the observations. However, a difficult problem that
arises is that observations made at a later time are invariably incompatible
with previously simulated states. To solve this problem, we propose a two-step
iterative process in which states of the system are alternately simulated in
accordance with past observations, then corrected in light of new observations.
We identify criteria under which such corrections can be made while maintaining
appropriate importance weights.
| stat |
A Metropolized adaptive subspace algorithm for high-dimensional Bayesian
variable selection | A simple and efficient adaptive Markov Chain Monte Carlo (MCMC) method,
called the Metropolized Adaptive Subspace (MAdaSub) algorithm, is proposed for
sampling from high-dimensional posterior model distributions in Bayesian
variable selection. The MAdaSub algorithm is based on an independent
Metropolis-Hastings sampler, where the individual proposal probabilities of the
explanatory variables are updated after each iteration using a form of Bayesian
adaptive learning, in a way that they finally converge to the respective
covariates' posterior inclusion probabilities. We prove the ergodicity of the
algorithm and present a parallel version of MAdaSub with an adaptation scheme
for the proposal probabilities based on the combination of information from
multiple chains. The effectiveness of the algorithm is demonstrated via various
simulated and real data examples, including a high-dimensional problem with
more than 20,000 covariates.
| stat |
A sub-critical branching process model for application to analysing Y
haplotype DNA mixtures | The treatment of short-tandem-repeat (STR) loci on the Y chromosome presents
special problems in the forensic analysis of DNA mixtures, chiefly but not
exclusively relating to the linkage of Y-STR loci which precludes the use of
the `product rule' for estimating Y-haplotype match probabilities. In recent
paper, Andersen and Balding(2017) estimated, via a population simulation model,
the distribution of the number of haplotypes sharing a common profile over a
set of Y-STR loci, and argued for its use as an alternative to estimating
Y-haplotype match probabilities.
In this paper we present a sub-critical branching process model that
approximates their population model, and show how to estimate the haplotype
number distribution numerically using multivariate probability generating
functions. It is shown that the approximation provides a good fit to their
simulations. The model is extended to propose a new framework for evaluating
the weight-of-evidence of Y-STR haplotype mixtures, and it is illustrated with
publicly available data of a three person DNA mixture.
| stat |
Relevance Vector Machine with Weakly Informative Hyperprior and Extended
Predictive Information Criterion | In the variational relevance vector machine, the gamma distribution is
representative as a hyperprior over the noise precision of automatic relevance
determination prior. Instead of the gamma hyperprior, we propose to use the
inverse gamma hyperprior with a shape parameter close to zero and a scale
parameter not necessary close to zero. This hyperprior is associated with the
concept of a weakly informative prior. The effect of this hyperprior is
investigated through regression to non-homogeneous data. Because it is
difficult to capture the structure of such data with a single kernel function,
we apply the multiple kernel method, in which multiple kernel functions with
different widths are arranged for input data. We confirm that the degrees of
freedom in a model is controlled by adjusting the scale parameter and keeping
the shape parameter close to zero. A candidate for selecting the scale
parameter is the predictive information criterion. However the estimated model
using this criterion seems to cause over-fitting. This is because the multiple
kernel method makes the model a situation where the dimension of the model is
larger than the data size. To select an appropriate scale parameter even in
such a situation, we also propose an extended prediction information criterion.
It is confirmed that a multiple kernel relevance vector regression model with
good predictive accuracy can be obtained by selecting the scale parameter
minimizing extended prediction information criterion.
| stat |
Fused-Lasso Regularized Cholesky Factors of Large Nonstationary
Covariance Matrices of Longitudinal Data | Smoothness of the subdiagonals of the Cholesky factor of large covariance
matrices is closely related to the degrees of nonstationarity of autoregressive
models for time series and longitudinal data. Heuristically, one expects for a
nearly stationary covariance matrix the entries in each subdiagonal of the
Cholesky factor of its inverse to be nearly the same in the sense that sum of
absolute values of successive terms is small. Statistically such smoothness is
achieved by regularizing each subdiagonal using fused-type lasso penalties. We
rely on the standard Cholesky factor as the new parameters within a regularized
normal likelihood setup which guarantees: (1) joint convexity of the likelihood
function, (2) strict convexity of the likelihood function restricted to each
subdiagonal even when $n<p$, and (3) positive-definiteness of the estimated
covariance matrix. A block coordinate descent algorithm, where each block is a
subdiagonal, is proposed and its convergence is established under mild
conditions. Lack of decoupling of the penalized likelihood function into a sum
of functions involving individual subdiagonals gives rise to some computational
challenges and advantages relative to two recent algorithms for sparse
estimation of the Cholesky factor which decouple row-wise. Simulation results
and real data analysis show the scope and good performance of the proposed
methodology.
| stat |
Information Mandala: Statistical Distance Matrix with Clustering | In machine learning, observation features are measured in a metric space to
obtain their distance function for optimization. Given similar features that
are statistically sufficient as a population, a statistical distance between
two probability distributions can be calculated for more precise learning.
Provided the observed features are multi-valued, the statistical distance
function is still efficient. However, due to its scalar output, it cannot be
applied to represent detailed distances between feature elements. To resolve
this problem, this paper extends the traditional statistical distance to a
matrix form, called a statistical distance matrix. In experiments, the proposed
approach performs well in object recognition tasks and clearly and intuitively
represents the dissimilarities between cat and dog images in the CIFAR dataset,
even when directly calculated using the image pixels. By using the hierarchical
clustering of the statistical distance matrix, the image pixels can be
separated into several clusters that are geometrically arranged around a center
like a Mandala pattern. The statistical distance matrix with clustering, called
the Information Mandala, is beyond ordinary saliency maps and can help to
understand the basic principles of the convolution neural network.
| stat |
An Introduction to Inductive Statistical Inference: from Parameter
Estimation to Decision-Making | These lecture notes aim at a post-Bachelor audience with a background at an
introductory level in Applied Mathematics and Applied Statistics. They discuss
the logic and methodology of the Bayes-Laplace approach to inductive
statistical inference that places common sense and the guiding lines of the
scientific method at the heart of systematic analyses of quantitative-empirical
data. Following an exposition of exactly solvable cases of single- and
two-parameter estimation problems, the main focus is laid on Markov Chain Monte
Carlo (MCMC) simulations on the basis of Hamiltonian Monte Carlo sampling of
posterior joint probability distributions for regression parameters occurring
in generalised linear models for a univariate outcome variable. The modelling
of fixed effects as well as of correlated varying effects via multi-level
models in non-centred parametrisation is considered. The simulation of
posterior predictive distributions is outlined. The assessment of a model's
relative out-of-sample posterior predictive accuracy with information
entropy-based criteria WAIC and LOOIC and model comparison with Bayes factors
are addressed. Concluding, a conceptual link to the behavioural subjective
expected utility representation of a single decision-maker's choice behaviour
in static one-shot decision problems is established. Vectorised codes for MCMC
simulations of multi-dimensional posterior joint probability distributions with
the Stan probabilistic programming language implemented in the statistical
software R are provided. The lecture notes are fully hyperlinked. They direct
the reader to original scientific research papers, online resources on
inductive statistical inference, and to pertinent biographical information.
| stat |
Deep Bayesian Nonparametric Factor Analysis | We propose a deep generative factor analysis model with beta process prior
that can approximate complex non-factorial distributions over the latent codes.
We outline a stochastic EM algorithm for scalable inference in a specific
instantiation of this model and present some preliminary results.
| stat |
More for less: Predicting and maximizing genetic variant discovery via
Bayesian nonparametrics | While the cost of sequencing genomes has decreased dramatically in recent
years, this expense often remains non-trivial. Under a fixed budget, then,
scientists face a natural trade-off between quantity and quality; they can
spend resources to sequence a greater number of genomes (quantity) or spend
resources to sequence genomes with increased accuracy (quality). Our goal is to
find the optimal allocation of resources between quantity and quality.
Optimizing resource allocation promises to reveal as many new variations in the
genome as possible, and thus as many new scientific insights as possible. In
this paper, we consider the common setting where scientists have already
conducted a pilot study to reveal variants in a genome and are contemplating a
follow-up study. We introduce a Bayesian nonparametric methodology to predict
the number of new variants in the follow-up study based on the pilot study.
When experimental conditions are kept constant between the pilot and follow-up,
we demonstrate on real data from the gnomAD project that our prediction is more
accurate than three recent proposals, and competitive with a more classic
proposal. Unlike existing methods, though, our method allows practitioners to
change experimental conditions between the pilot and the follow-up. We
demonstrate how this distinction allows our method to be used for (i) more
realistic predictions and (ii) optimal allocation of a fixed budget between
quality and quantity.
| stat |
Conditional Inferences Based on Vine Copulas with Applications to Credit
Spread Data of Corporate Bonds | Understanding the dependence relationship of credit spreads of corporate
bonds is important for risk management. Vine copula models with tail dependence
are used to analyze a credit spread dataset of Chinese corporate bonds,
understand the dependence among different sectors and perform conditional
inferences. It is shown how the effect of tail dependence affects risk
transfer, or the conditional distributions given one variable is extreme. Vine
copula models also provide more accurate cross prediction results compared with
linear regressions. These conditional inference techniques are a statistical
contribution for analysis of bond credit spreads of investment portfolios
consisting of corporate bonds from various sectors.
| stat |
Hard and Soft EM in Bayesian Network Learning from Incomplete Data | Incomplete data are a common feature in many domains, from clinical trials to
industrial applications. Bayesian networks (BNs) are often used in these
domains because of their graphical and causal interpretations. BN parameter
learning from incomplete data is usually implemented with the
Expectation-Maximisation algorithm (EM), which computes the relevant sufficient
statistics ("soft EM") using belief propagation. Similarly, the Structural
Expectation-Maximisation algorithm (Structural EM) learns the network structure
of the BN from those sufficient statistics using algorithms designed for
complete data. However, practical implementations of parameter and structure
learning often impute missing data ("hard EM") to compute sufficient statistics
instead of using belief propagation, for both ease of implementation and
computational speed. In this paper, we investigate the question: what is the
impact of using imputation instead of belief propagation on the quality of the
resulting BNs? From a simulation study using synthetic data and reference BNs,
we find that it is possible to recommend one approach over the other in several
scenarios based on the characteristics of the data. We then use this
information to build a simple decision tree to guide practitioners in choosing
the EM algorithm best suited to their problem.
| stat |
Error bounds in estimating the out-of-sample prediction error using
leave-one-out cross validation in high-dimensions | We study the problem of out-of-sample risk estimation in the high dimensional
regime where both the sample size $n$ and number of features $p$ are large, and
$n/p$ can be less than one. Extensive empirical evidence confirms the accuracy
of leave-one-out cross validation (LO) for out-of-sample risk estimation. Yet,
a unifying theoretical evaluation of the accuracy of LO in high-dimensional
problems has remained an open problem. This paper aims to fill this gap for
penalized regression in the generalized linear family. With minor assumptions
about the data generating process, and without any sparsity assumptions on the
regression coefficients, our theoretical analysis obtains finite sample upper
bounds on the expected squared error of LO in estimating the out-of-sample
error. Our bounds show that the error goes to zero as $n,p \rightarrow \infty$,
even when the dimension $p$ of the feature vectors is comparable with or
greater than the sample size $n$. One technical advantage of the theory is that
it can be used to clarify and connect some results from the recent literature
on scalable approximate LO.
| stat |
not-MIWAE: Deep Generative Modelling with Missing not at Random Data | When a missing process depends on the missing values themselves, it needs to
be explicitly modelled and taken into account while doing likelihood-based
inference. We present an approach for building and fitting deep latent variable
models (DLVMs) in cases where the missing process is dependent on the missing
data. Specifically, a deep neural network enables us to flexibly model the
conditional distribution of the missingness pattern given the data. This allows
for incorporating prior information about the type of missingness (e.g.
self-censoring) into the model. Our inference technique, based on
importance-weighted variational inference, involves maximising a lower bound of
the joint likelihood. Stochastic gradients of the bound are obtained by using
the reparameterisation trick both in latent space and data space. We show on
various kinds of data sets and missingness patterns that explicitly modelling
the missing process can be invaluable.
| stat |
Unified Rules of Renewable Weighted Sums for Various Online Updating
Estimations | This paper establishes unified frameworks of renewable weighted sums (RWS)
for various online updating estimations in the models with streaming data sets.
The newly defined RWS lays the foundation of online updating likelihood, online
updating loss function, online updating estimating equation and so on. The idea
of RWS is intuitive and heuristic, and the algorithm is computationally simple.
This paper chooses nonparametric model as an exemplary setting. The RWS applies
to various types of nonparametric estimators, which include but are not limited
to nonparametric likelihood, quasi-likelihood and least squares. Furthermore,
the method and the theory can be extended into the models with both parameter
and nonparametric function. The estimation consistency and asymptotic normality
of the proposed renewable estimator are established, and the oracle property is
obtained. Moreover, these properties are always satisfied, without any
constraint on the number of data batches, which means that the new method is
adaptive to the situation where streaming data sets arrive perpetually. The
behavior of the method is further illustrated by various numerical examples
from simulation experiments and real data analysis.
| stat |
Power prior models for treatment effect estimation in a small n,
sequential, multiple assignment, randomized trial | A small n, sequential, multiple assignment, randomized trial (snSMART) is a
small sample, two-stage design where participants receive up to two treatments
sequentially, but the second treatment depends on response to the first
treatment. The treatment effect of interest in an snSMART is the first-stage
response rate, but outcomes from both stages can be used to obtain more
information from a small sample. A novel way to incorporate the outcomes from
both stages applies power prior models, in which first stage outcomes from an
snSMART are regarded as the primary data and second stage outcomes are regarded
as supplemental. We apply existing power prior models to snSMART data, and we
also develop new extensions of power prior models. All methods are compared to
each other and to the Bayesian joint stage model (BJSM) via simulation studies.
By comparing the biases and the efficiency of the response rate estimates among
all proposed power prior methods, we suggest application of Fisher's exact test
or the Bhattacharyya's overlap measure to an snSMART to estimate the treatment
effect in an snSMART, which both have performance mostly as good or better than
the BJSM. We describe the situations where each of these suggested approaches
is preferred.
| stat |
Benchmarking Minimax Linkage | Minimax linkage was first introduced by Ao et al. [3] in 2004, as an
alternative to standard linkage methods used in hierarchical clustering.
Minimax linkage relies on distances to a prototype for each cluster; this
prototype can be thought of as a representative object in the cluster, hence
improving the interpretability of clustering results. Bien and Tibshirani
analyzed properties of this method in 2011 [2], popularizing the method within
the statistics community. Additionally, they performed comparisons of minimax
linkage to standard linkage methods, making use of five data sets and two
different evaluation metrics (distance to prototype and misclassification
rate). In an effort to expand upon their work and evaluate minimax linkage more
comprehensively, our benchmark study focuses on thorough method evaluation via
multiple performance metrics on several well-described data sets. We also make
all code and data publicly available through an R package, for full
reproducibility. Similarly to [2], we find that minimax linkage often produces
the smallest maximum minimax radius of all linkage methods, meaning that
minimax linkage produces clusters where objects in a cluster are tightly
clustered around their prototype. This is true across a range of values for the
total number of clusters (k). However, this is not always the case, and special
attention should be paid to the case when k is the true known value. For true
k, minimax linkage does not always perform the best in terms of all the
evaluation metrics studied, including maximum minimax radius. This paper was
motivated by the IFCS Cluster Benchmarking Task Force's call for clustering
benchmark studies and the white paper [5], which put forth guidelines and
principles for comprehensive benchmarking in clustering. Our work is designed
to be a neutral benchmark study of minimax linkage.
| stat |
Using an online sample to learn about an offline population | Online data sources offer tremendous promise to demography and other social
sciences, but researchers worry that the group of people who are represented in
online datasets can be different from the general population. We show that by
sampling and anonymously interviewing people who are online, researchers can
learn about both people who are online and people who are offline. Our approach
is based on the insight that people everywhere are connected through in-person
social networks, such as kin, friendship, and contact networks. We illustrate
how this insight can be used to derive an estimator for tracking the *digital
divide* in access to the internet, an increasingly important dimension of
population inequality in the modern world. We conducted a large-scale empirical
test of our approach, using an online sample to estimate internet adoption in
five countries ($n \approx 15,000$). Our test embedded a randomized experiment
whose results can help design future studies. Our approach could be adapted to
many other settings, offering one way to overcome some of the major challenges
facing demographers in the information age.
| stat |
Orthogonal Impulse Response Analysis in Presence of Time-Varying
Covariance | In this paper the orthogonal impulse response functions (OIRF) are studied in
the non-standard, though quite common, case where the covariance of the error
vector is not constant in time. The usual approach for taking into account such
behavior of the covariance consists in applying the standard tools to
sub-periods of the whole sample. We underline that such a practice may lead to
severe upward bias. We propose a new approach intended to give what we argue to
be a more accurate resume of the time-varying OIRF. This consists in averaging
the Cholesky decomposition of nonparametric covariance estimators. In addition
an index is developed to evaluate the heteroscedasticity effect on the OIRF
analysis. The asymptotic behavior of the different estimators considered in the
paper is investigated. The theoretical results are illustrated by Monte Carlo
experiments. The analysis of the orthogonal response functions of the U.S.
inflation to an oil price shock, shows the relevance of the tools proposed
herein for an appropriate analysis of economic variables.
| stat |
Degrees of freedom for off-the-grid sparse estimation | A central question in modern machine learning and imaging sciences is to
quantify the number of effective parameters of vastly over-parameterized
models. The degrees of freedom is a mathematically convenient way to define
this number of parameters. Its computation and properties are well understood
when dealing with discretized linear models, possibly regularized using
sparsity. In this paper, we argue that this way of thinking is plagued when
dealing with models having very large parameter spaces. In this case it makes
more sense to consider "off-the-grid" approaches, using a continuous parameter
space. This type of approach is the one favoured when training multi-layer
perceptrons, and is also becoming popular to solve super-resolution problems in
imaging. Training these off-the-grid models with a sparsity inducing prior can
be achieved by solving a convex optimization problem over the space of
measures, which is often called the Beurling Lasso (Blasso), and is the
continuous counterpart of the celebrated Lasso parameter selection method. In
previous works, the degrees of freedom for the Lasso was shown to coincide with
the size of the smallest solution support. Our main contribution is a proof of
a continuous counterpart to this result for the Blasso. Our findings suggest
that discretized methods actually vastly over-estimate the number of intrinsic
continuous degrees of freedom. Our second contribution is a detailed study of
the case of sampling Fourier coefficients in 1D, which corresponds to a
super-resolution problem. We show that our formula for the degrees of freedom
is valid outside of a set of measure zero of observations, which in turn
justifies its use to compute an unbiased estimator of the prediction risk using
the Stein Unbiased Risk Estimator (SURE).
| stat |
Chain effects of clean water: The Mills-Reincke phenomenon in early
twentieth-century Japan | This study explores the validity of chain effects of clean water, which are
known as the "Mills-Reincke phenomenon," in early twentieth-century Japan.
Recent studies have reported that water purifications systems are responsible
for huge contributions to human capital. Although some studies have
investigated the instantaneous effects of water-supply systems in pre-war
Japan, little is known about the chain effects of these systems. By analyzing
city-level cause-specific mortality data from 1922-1940, we find that a decline
in typhoid deaths by one per 1,000 people decreased the risk of death due to
non-waterborne diseases such as tuberculosis and pneumonia by 0.742-2.942 per
1,000 people. Our finding suggests that the observed Mills-Reincke phenomenon
could have resulted in the relatively rapid decline in the mortality rate in
early twentieth-century Japan.
| stat |
Sparse Representation Classification Beyond L1 Minimization and the
Subspace Assumption | The sparse representation classifier (SRC) has been utilized in various
classification problems, which makes use of L1 minimization and works well for
image recognition satisfying a subspace assumption. In this paper we propose a
new implementation of SRC via screening, establish its equivalence to the
original SRC under regularity conditions, and prove its classification
consistency under a latent subspace model and contamination. The results are
demonstrated via simulations and real data experiments, where the new algorithm
achieves comparable numerical performance and significantly faster.
| stat |
New plans orthogonal through the block factor | In the present paper we construct plans orthogonal through the block factor
(POTBs). We describe procedures for adding blocks as well as factors to an
initial plan and thus generate a bigger plan. Using these procedures we
construct POTBs for symmetrical experiments with factors having three or more
levels. We also construct a series of plans inter-class orthogonal through the
block factor for two-level factors.
| stat |
Stochastic Comparison of Parallel Systems with Log-Lindley Distributed
Components under Random Shocks | Recently, Chowdhury and Kundu [6] compared two parallel systems of
heterogeneous independent log-Lindley distributed components using the concept
of vector majorization and related orders. Under the same set-up, this paper
derives some results related to usual stochastic ordering between two parallel
systems when each component receives a random shock.
| stat |
BOLT-SSI: A Statistical Approach to Screening Interaction Effects for
Ultra-High Dimensional Data | Detecting interaction effects among predictors on the response variable is a
crucial step in various applications. In this paper, we first propose a simple
method for sure screening interactions (SSI). Although its computation
complexity is $O(p^2n)$, SSI works well for problems of moderate dimensionality
(e.g., $p=10^3\sim10^4$), without the heredity assumption. To ultra-high
dimensional problems (e.g., $p = 10^6$), motivated by discretization associated
Boolean representation and operations and the contingency table for discrete
variables, we propose a fast algorithm, named "BOLT-SSI". The statistical
theory has been established for SSI and BOLT-SSI, guaranteeing their sure
screening property. The performance of SSI and BOLT-SSI are evaluated by
comprehensive simulation and real case studies. Numerical results demonstrate
that SSI and BOLT-SSI can often outperform their competitors in terms of
computational efficiency and statistical accuracy. The proposed method can be
applied for fully detecting interactions with more than 300,000 predictors.
Based on this study, we believe that there is a great need to rethink the
relationship between statistical accuracy and computational efficiency. We have
shown that the computational performance of a statistical method can often be
greatly improved by exploring the advantages of computational architecture with
a tolerable loss of statistical accuracy.
| stat |
Limitations of Pinned AUC for Measuring Unintended Bias | This report examines the Pinned AUC metric introduced and highlights some of
its limitations. Pinned AUC provides a threshold-agnostic measure of unintended
bias in a classification model, inspired by the ROC-AUC metric. However, as we
highlight in this report, there are ways that the metric can obscure different
kinds of unintended biases when the underlying class distributions on which
bias is being measured are not carefully controlled.
| stat |
Expected Policy Gradients for Reinforcement Learning | We propose expected policy gradients (EPG), which unify stochastic policy
gradients (SPG) and deterministic policy gradients (DPG) for reinforcement
learning. Inspired by expected sarsa, EPG integrates (or sums) across actions
when estimating the gradient, instead of relying only on the action in the
sampled trajectory. For continuous action spaces, we first derive a practical
result for Gaussian policies and quadratic critics and then extend it to a
universal analytical method, covering a broad class of actors and critics,
including Gaussian, exponential families, and policies with bounded support.
For Gaussian policies, we introduce an exploration method that uses covariance
proportional to the matrix exponential of the scaled Hessian of the critic with
respect to the actions. For discrete action spaces, we derive a variant of EPG
based on softmax policies. We also establish a new general policy gradient
theorem, of which the stochastic and deterministic policy gradient theorems are
special cases. Furthermore, we prove that EPG reduces the variance of the
gradient estimates without requiring deterministic policies and with little
computational overhead. Finally, we provide an extensive experimental
evaluation of EPG and show that it outperforms existing approaches on multiple
challenging control domains.
| stat |
The Medical Deconfounder: Assessing Treatment Effects with Electronic
Health Records | The treatment effects of medications play a key role in guiding medical
prescriptions. They are usually assessed with randomized controlled trials
(RCTs), which are expensive. Recently, large-scale electronic health records
(EHRs) have become available, opening up new opportunities for more
cost-effective assessments. However, assessing a treatment effect from EHRs is
challenging: it is biased by unobserved confounders, unmeasured variables that
affect both patients' medical prescription and their outcome, e.g. the
patients' social economic status. To adjust for unobserved confounders, we
develop the medical deconfounder, a machine learning algorithm that unbiasedly
estimates treatment effects from EHRs. The medical deconfounder first
constructs a substitute confounder by modeling which medications were
prescribed to each patient; this substitute confounder is guaranteed to capture
all multi-medication confounders, observed or unobserved (arXiv:1805.06826). It
then uses this substitute confounder to adjust for the confounding bias in the
analysis. We validate the medical deconfounder on two simulated and two real
medical data sets. Compared to classical approaches, the medical deconfounder
produces closer-to-truth treatment effect estimates; it also identifies
effective medications that are more consistent with the findings in the medical
literature.
| stat |
Localization for MCMC: sampling high-dimensional posterior distributions
with local structure | We investigate how ideas from covariance localization in numerical weather
prediction can be used in Markov chain Monte Carlo (MCMC) sampling of
high-dimensional posterior distributions arising in Bayesian inverse problems.
To localize an inverse problem is to enforce an anticipated "local" structure
by (i) neglecting small off-diagonal elements of the prior precision and
covariance matrices; and (ii) restricting the influence of observations to
their neighborhood. For linear problems we can specify the conditions under
which posterior moments of the localized problem are close to those of the
original problem. We explain physical interpretations of our assumptions about
local structure and discuss the notion of high dimensionality in local
problems, which is different from the usual notion of high dimensionality in
function space MCMC. The Gibbs sampler is a natural choice of MCMC algorithm
for localized inverse problems and we demonstrate that its convergence rate is
independent of dimension for localized linear problems. Nonlinear problems can
also be tackled efficiently by localization and, as a simple illustration of
these ideas, we present a localized Metropolis-within-Gibbs sampler. Several
linear and nonlinear numerical examples illustrate localization in the context
of MCMC samplers for inverse problems.
| stat |
Sparse Projection Oblique Randomer Forests | Decision forests, including Random Forests and Gradient Boosting Trees, have
recently demonstrated state-of-the-art performance in a variety of machine
learning settings. Decision forests are typically ensembles of axis-aligned
decision trees; that is, trees that split only along feature dimensions. In
contrast, many recent extensions to decision forests are based on axis-oblique
splits. Unfortunately, these extensions forfeit one or more of the favorable
properties of decision forests based on axis-aligned splits, such as robustness
to many noise dimensions, interpretability, or computational efficiency. We
introduce yet another decision forest, called "Sparse Projection Oblique
Randomer Forests" (SPORF). SPORF uses very sparse random projections, i.e.,
linear combinations of a small subset of features. SPORF significantly improves
accuracy over existing state-of-the-art algorithms on a standard benchmark
suite for classification with >100 problems of varying dimension, sample size,
and number of classes. To illustrate how SPORF addresses the limitations of
both axis-aligned and existing oblique decision forest methods, we conduct
extensive simulated experiments. SPORF typically yields improved performance
over existing decision forests, while mitigating computational efficiency and
scalability and maintaining interpretability. SPORF can easily be incorporated
into other ensemble methods such as boosting to obtain potentially similar
gains.
| stat |
Human Action Attribute Learning From Video Data Using Low-Rank
Representations | Representation of human actions as a sequence of human body movements or
action attributes enables the development of models for human activity
recognition and summarization. We present an extension of the low-rank
representation (LRR) model, termed the clustering-aware structure-constrained
low-rank representation (CS-LRR) model, for unsupervised learning of human
action attributes from video data. Our model is based on the union-of-subspaces
(UoS) framework, and integrates spectral clustering into the LRR optimization
problem for better subspace clustering results. We lay out an efficient linear
alternating direction method to solve the CS-LRR optimization problem. We also
introduce a hierarchical subspace clustering approach, termed hierarchical
CS-LRR, to learn the attributes without the need for a priori specification of
their number. By visualizing and labeling these action attributes, the
hierarchical model can be used to semantically summarize long video sequences
of human actions at multiple resolutions. A human action or activity can also
be uniquely represented as a sequence of transitions from one action attribute
to another, which can then be used for human action recognition. We demonstrate
the effectiveness of the proposed model for semantic summarization and action
recognition through comprehensive experiments on five real-world human action
datasets.
| stat |
High-Dimensional Changepoint Detection via a Geometrically Inspired
Mapping | High-dimensional changepoint analysis is a growing area of research and has
applications in a wide range of fields. The aim is to accurately and
efficiently detect changepoints in time series data when both the number of
time points and dimensions grow large. Existing methods typically aggregate or
project the data to a smaller number of dimensions; usually one. We present a
high-dimensional changepoint detection method that takes inspiration from
geometry to map a high-dimensional time series to two dimensions. We show
theoretically and through simulation that if the input series is Gaussian then
the mappings preserve the Gaussianity of the data. Applying univariate
changepoint detection methods to both mapped series allows the detection of
changepoints that correspond to changes in the mean and variance of the
original time series. We demonstrate that this approach outperforms the current
state-of-the-art multivariate changepoint methods in terms of accuracy of
detected changepoints and computational efficiency. We conclude with
applications from genetics and finance.
| stat |
Tolerance and Prediction Intervals for Non-normal Models | A prediction interval covers a future observation from a random process in
repeated sampling, and is typically constructed by identifying a pivotal
quantity that is also an ancillary statistic. Analogously, a tolerance interval
covers a population percentile in repeated sampling and is often based on a
pivotal quantity. One approach we consider in non-normal models leverages a
link function resulting in a pivotal quantity that is approximately normally
distributed. In settings where this normal approximation does not hold we
consider a second approach for tolerance and prediction based on a confidence
interval for the mean. These methods are intuitive, simple to implement, have
proper operating characteristics, and are computationally efficient compared to
Bayesian, re-sampling, and machine learning methods. This is demonstrated in
the context of multi-site clinical trial recruitment with staggered site
initiation, real-world time on treatment, and end-of-study success for a
clinical endpoint.
| stat |
Forecasting reconciliation with a top-down alignment of independent
level forecasts | Hierarchical forecasting with intermittent time series is a challenge in both
research and empirical studies. The overall forecasting performance is heavily
affected by the forecasting accuracy of intermittent time series at bottom
levels. In this paper, we present a forecasting reconciliation approach that
treats the bottom level forecast as latent to ensure higher forecasting
accuracy on the upper levels of the hierarchy. We employ a pure deep learning
forecasting approach N-BEATS for continuous time series on top levels and a
widely used tree-based algorithm LightGBM for the bottom level intermittent
time series. The hierarchical forecasting with alignment approach is simple and
straightforward to implement in practice. It sheds light on an orthogonal
direction for forecasting reconciliation. When there is difficulty finding an
optimal reconciliation, allowing suboptimal forecasts at a lower level could
retain a high overall performance. The approach in this empirical study was
developed by the first author during the M5 Forecasting Accuracy competition
ranking second place. The approach is business orientated and could be
beneficial for business strategic planning.
| stat |
Latent Network Structure Learning from High Dimensional Multivariate
Point Processes | Learning the latent network structure from large scale multivariate point
process data is an important task in a wide range of scientific and business
applications. For instance, we might wish to estimate the neuronal functional
connectivity network based on spiking times recorded from a collection of
neurons. To characterize the complex processes underlying the observed data, we
propose a new and flexible class of nonstationary Hawkes processes that allow
both excitatory and inhibitory effects. We estimate the latent network
structure using an efficient sparse least squares estimation approach. Using a
thinning representation, we establish concentration inequalities for the first
and second order statistics of the proposed Hawkes process. Such theoretical
results enable us to establish the non-asymptotic error bound and the selection
consistency of the estimated parameters. Furthermore, we describe a least
squares loss based statistic for testing if the background intensity is
constant in time. We demonstrate the efficacy of our proposed method through
simulation studies and an application to a neuron spike train data set.
| stat |
Density Estimation by Monte Carlo and Quasi-Monte Carlo | Estimating the density of a continuous random variable X has been studied
extensively in statistics, in the setting where n independent observations of X
are given a priori and one wishes to estimate the density from that. Popular
methods include histograms and kernel density estimators. In this review paper,
we are interested instead in the situation where the observations are generated
by Monte Carlo simulation from a model. Then, one can take advantage of
variance reduction methods such as stratification, conditional Monte Carlo, and
randomized quasi-Monte Carlo (RQMC), and obtain a more accurate density
estimator than with standard Monte Carlo for a given computing budget. We
discuss several ways of doing this, proposed in recent papers, with a focus on
methods that exploit RQMC. A first idea is to directly combine RQMC with a
standard kernel density estimator. Another one is to adapt a simulation-based
derivative estimation method such as smoothed perturbation analysis or the
likelihood ratio method to obtain a continuous estimator of the cdf, whose
derivative is an unbiased estimator of the density. This can then be combined
with RQMC. We summarize recent theoretical results with these approaches and
give numerical illustrations of how they improve the convergence of the mean
square integrated error.
| stat |
Semiparametric Wavelet-based JPEG IV Estimator for endogenously
truncated data | A new and an enriched JPEG algorithm is provided for identifying redundancies
in a sequence of irregular noisy data points which also accommodates a
reference-free criterion function. Our main contribution is by formulating
analytically (instead of approximating) the inverse of the transpose of
JPEGwavelet transform without involving matrices which are computationally
cumbersome. The algorithm is suitable for the widely-spread situations where
the original data distribution is unobservable such as in cases where there is
deficient representation of the entire population in the training data (in
machine learning) and thus the covariate shift assumption is violated. The
proposed estimator corrects for both biases, the one generated by endogenous
truncation and the one generated by endogenous covariates. Results from
utilizing 2,000,000 different distribution functions verify the applicability
and high accuracy of our procedure to cases in which the disturbances are
neither jointly nor marginally normally distributed.
| stat |
A Projector-Based Approach to Quantifying Total and Excess Uncertainties
for Sketched Linear Regression | Linear regression is a classic method of data analysis. In recent years,
sketching -- a method of dimension reduction using random sampling, random
projections, or both -- has gained popularity as an effective computational
approximation when the number of observations greatly exceeds the number of
variables. In this paper, we address the following question: How does sketching
affect the statistical properties of the solution and key quantities derived
from it?
To answer this question, we present a projector-based approach to sketched
linear regression that is exact and that requires minimal assumptions on the
sketching matrix. Therefore, downstream analyses hold exactly and generally for
all sketching schemes. Additionally, a projector-based approach enables
derivation of key quantities from classic linear regression that account for
the combined model- and algorithm-induced uncertainties. We demonstrate the
usefulness of a projector-based approach in quantifying and enabling insight on
excess uncertainties and bias-variance decompositions for sketched linear
regression. Finally, we demonstrate how the insights from our projector-based
analyses can be used to produce practical sketching diagnostics to aid the
design of judicious sketching schemes.
| stat |
Principal Stratum Strategy: Potential Role in Drug Development | A randomized trial allows estimation of the causal effect of an intervention
compared to a control in the overall population and in subpopulations defined
by baseline characteristics. Often, however, clinical questions also arise
regarding the treatment effect in subpopulations of patients, which would
experience clinical or disease related events post-randomization. Events that
occur after treatment initiation and potentially affect the interpretation or
the existence of the measurements are called {\it intercurrent events} in the
ICH E9(R1) guideline. If the intercurrent event is a consequence of treatment,
randomization alone is no longer sufficient to meaningfully estimate the
treatment effect. Analyses comparing the subgroups of patients without the
intercurrent events for intervention and control will not estimate a causal
effect. This is well known, but post-hoc analyses of this kind are commonly
performed in drug development. An alternative approach is the principal stratum
strategy, which classifies subjects according to their potential occurrence of
an intercurrent event on both study arms. We illustrate with examples that
questions formulated through principal strata occur naturally in drug
development and argue that approaching these questions with the ICH E9(R1)
estimand framework has the potential to lead to more transparent assumptions as
well as more adequate analyses and conclusions. In addition, we provide an
overview of assumptions required for estimation of effects in principal strata.
Most of these assumptions are unverifiable and should hence be based on solid
scientific understanding. Sensitivity analyses are needed to assess robustness
of conclusions.
| stat |
Student Log-Data from a Randomized Evaluation of Educational Technology:
A Causal Case Study | Randomized evaluations of educational technology produce log data as a
bi-product: highly granular data student and teacher usage. These datasets
could shed light on causal mechanisms, effect heterogeneity, or optimal use.
However, there are methodological challenges: implementation is not randomized
and is only defined for the treatment group, and log datasets have a complex
structure. This paper discusses three approaches to help surmount these issues.
One approach uses data from the treatment group to estimate the effect of usage
on outcomes in an observational study. Another, causal mediation analysis,
estimates the role of usage in driving the overall effect. Finally, principal
stratification estimates overall effects for groups of students with the same
"potential" usage. We analyze hint data from an evaluation of the Cognitive
Tutor Algebra I curriculum using these three approaches, with possibly
conflicting results: the observational study and mediation analysis suggest
that hints reduce posttest scores, while principal stratification finds that
treatment effects may be correlated with higher rates of hint requests. We
discuss these mixed conclusions and give broader methodological
recommendations.
| stat |
Improved Convergence Speed of Fully Symmetric Learning Rules for
Principal Component Analysis | Fully symmetric learning rules for principal component analysis can be
derived from a novel objective function suggested in our previous work. We
observed that these learning rules suffer from slow convergence for covariance
matrices where some principal eigenvalues are close to each other. Here we
describe a modified objective function with an additional term which mitigates
this convergence problem. We show that the learning rule derived from the
modified objective function inherits all fixed points from the original
learning rule (but may introduce additional ones). Also the stability of the
inherited fixed points remains unchanged. Only the steepness of the objective
function is increased in some directions. Simulations confirm that the
convergence speed can be noticeably improved, depending on the weight factor of
the additional term.
| stat |
An Efficient Semi-smooth Newton Augmented Lagrangian Method for Elastic
Net | Feature selection is an important and active research area in statistics and
machine learning. The Elastic Net is often used to perform selection when the
features present non-negligible collinearity or practitioners wish to
incorporate additional known structure. In this article, we propose a new
Semi-smooth Newton Augmented Lagrangian Method to efficiently solve the Elastic
Net in ultra-high dimensional settings. Our new algorithm exploits both the
sparsity induced by the Elastic Net penalty and the sparsity due to the second
order information of the augmented Lagrangian. This greatly reduces the
computational cost of the problem. Using simulations on both synthetic and real
datasets, we demonstrate that our approach outperforms its best competitors by
at least an order of magnitude in terms of CPU time. We also apply our approach
to a Genome Wide Association Study on childhood obesity.
| stat |
"Predicting" after peeking into the future: Correcting a fundamental
flaw in the SAOM -- TERGM comparison of Leifeld and Cranmer (2019) | We review the empirical comparison of SAOMs and TERGMs by Leifeld and Cranmer
(2019) in Network Science. We note that their model specification uses nodal
covariates calculated from observed degrees instead of using structural
effects, thus turning endogeneity into circularity. In consequence, their
out-of-sample predictions using TERGMs are based on out-of-sample information
and thereby predict the future using observations from the future. We conclude
that their analysis rest on erroneous model specifications that render the
article's conclusions meaningless. Consequently, researchers should disregard
recommendations from the criticized paper when making informed modelling
choices.
| stat |
Cross-temporal forecast reconciliation: Optimal combination method and
heuristic alternatives | Forecast reconciliation is a post-forecasting process aimed to improve the
quality of the base forecasts for a system of hierarchical/grouped time series
(Hyndman et al., 2011). Contemporaneous (cross-sectional) and temporal
hierarchies have been considered in the literature, but - except for Kourentzes
and Athanasopoulos (2019) - generally these two features have not been fully
considered together. Adopting a notation able to simultaneously deal with both
forecast reconciliation dimensions, the paper shows two new results: (i) an
iterative cross-temporal forecast reconciliation procedure which extends, and
overcomes some weaknesses of, the two-step procedure by Kourentzes and
Athanasopoulos (2019), and (ii) the closed-form expression of the optimal (in
least squares sense) point forecasts which fulfill both contemporaneous and
temporal constraints. The feasibility of the proposed procedures, along with
first evaluations of their performance as compared to the most performing
`single dimension' (either cross-sectional or temporal) forecast reconciliation
procedures, is studied through a forecasting experiment on the 95 quarterly
time series of the Australian GDP from Income and Expenditure sides considered
by Athanasopoulos et al. (2019).
| stat |
New statistic for detecting laboratory effects in ORDANOVA | The present study defines a new statistic for detecting laboratory effects in
the analysis of ordinal variation (ORDANOVA). The ORDANOVA is an analysis
method similar to one-way analysis of variance for analysing ordinal data
obtained from interlaboratory comparison studies. In this paper, we present an
approximate continuous distribution for the new statistic for the case of an
arbitrary number of ordinal levels, and we demonstrate that $alpha$-percentiles
of the distribution are suitable criteria for conducting statistical tests. In
addition, a real example involving data from an interlaboratory comparison
study is analysed using the proposed statistic.
| stat |
Accelerating Experimental Design by Incorporating Experimenter Hunches | Experimental design is a process of obtaining a product with target property
via experimentation. Bayesian optimization offers a sample-efficient tool for
experimental design when experiments are expensive. Often, expert experimenters
have 'hunches' about the behavior of the experimental system, offering
potentials to further improve the efficiency. In this paper, we consider
per-variable monotonic trend in the underlying property that results in a
unimodal trend in those variables for a target value optimization. For example,
sweetness of a candy is monotonic to the sugar content. However, to obtain a
target sweetness, the utility of the sugar content becomes a unimodal function,
which peaks at the value giving the target sweetness and falls off both ways.
In this paper, we propose a novel method to solve such problems that achieves
two main objectives: a) the monotonicity information is used to the fullest
extent possible, whilst ensuring that b) the convergence guarantee remains
intact. This is achieved by a two-stage Gaussian process modeling, where the
first stage uses the monotonicity trend to model the underlying property, and
the second stage uses `virtual' samples, sampled from the first, to model the
target value optimization function. The process is made theoretically
consistent by adding appropriate adjustment factor in the posterior
computation, necessitated because of using the `virtual' samples. The proposed
method is evaluated through both simulations and real world experimental design
problems of a) new short polymer fiber with the target length, and b) designing
of a new three dimensional porous scaffolding with a target porosity. In all
scenarios our method demonstrates faster convergence than the basic Bayesian
optimization approach not using such `hunches'.
| stat |
Semiparametric Regression for Dual Population Mortality | Parameter shrinkage applied optimally can always reduce error and projection
variances from those of maximum likelihood estimation. Many variables that
actuaries use are on numerical scales, like age or year, which require
parameters at each point. Rather than shrinking these towards zero, nearby
parameters are better shrunk towards each other. Semiparametric regression is a
statistical discipline for building curves across parameter classes using
shrinkage methodology. It is similar to but more parsimonious than cubic
splines. We introduce it in the context of Bayesian shrinkage and apply it to
joint mortality modeling for related populations. Bayesian shrinkage of slope
changes of linear splines is an approach to semiparametric modeling that
evolved in the actuarial literature. It has some theoretical and practical
advantages, like closed-form curves, direct and transparent determination of
degree of shrinkage and of placing knots for the splines, and quantifying
goodness of fit. It is also relatively easy to apply to the many nonlinear
models that arise in actuarial work. We find that it compares well to a more
complex state-of-the-art statistical spline shrinkage approach on a popular
example from that literature.
| stat |
merlin: An R package for Mixed Effects Regression for Linear, Nonlinear
and User-defined models | The R package merlin performs flexible joint modelling of hierarchical
multi-outcome data. Increasingly, multiple longitudinal biomarker measurements,
possibly censored time-to-event outcomes and baseline characteristics are
available. However, there is limited software that allows all of this
information to be incorporated into one model. In this paper, we present merlin
which allows for the estimation of models with unlimited numbers of continuous,
binary, count and time-to-event outcomes, with unlimited levels of nested
random effects. A wide variety of link functions, including the expected value,
the gradient and shared random effects, are available in order to link the
different outcomes in a biologically plausible way. The accompanying
predict.merlin function allows for individual and population level predictions
to be made from even the most complex models. There is the option to specify
user-defined families, making merlin ideal for methodological research. The
flexibility of merlin is illustrated using an example in patients followed up
after heart valve replacement, beginning with a linear model, and finishing
with a joint multiple longitudinal and competing risks survival model.
| stat |
Data-driven confidence bands for distributed nonparametric regression | Gaussian Process Regression and Kernel Ridge Regression are popular
nonparametric regression approaches. Unfortunately, they suffer from high
computational complexity rendering them inapplicable to the modern massive
datasets. To that end a number of approximations have been suggested, some of
them allowing for a distributed implementation. One of them is the divide and
conquer approach, splitting the data into a number of partitions, obtaining the
local estimates and finally averaging them. In this paper we suggest a novel
computationally efficient fully data-driven algorithm, quantifying uncertainty
of this method, yielding frequentist $L_2$-confidence bands. We rigorously
demonstrate validity of the algorithm. Another contribution of the paper is a
minimax-optimal high-probability bound for the averaged estimator,
complementing and generalizing the known risk bounds.
| stat |
A computationally efficient neural network for predicting weather
forecast probabilities | The success of deep learning techniques over the last decades has opened up a
new avenue of research for weather forecasting. Here, we take the novel
approach of using a neural network to predict probability density functions
rather than a single output value, thus producing a probabilistic weather
forecast. This enables the calculation of both uncertainty and skill metrics
for the neural network predictions, and overcomes the common difficulty of
inferring uncertainty from these predictions.
This approach is purely data-driven and the neural network is trained on the
WeatherBench dataset (processed ERA5 data) to forecast geopotential and
temperature 3 and 5 days ahead. An extensive data exploration leads to the
identification of the most important input variables, which are also found to
agree with physical reasoning, thereby validating our approach. In order to
increase computational efficiency further, each neural network is trained on a
small subset of these variables. The outputs are then combined through a
stacked neural network, the first time such a technique has been applied to
weather data. Our approach is found to be more accurate than some numerical
weather prediction models and as accurate as more complex alternative neural
networks, with the added benefit of providing key probabilistic information
necessary for making informed weather forecasts.
| stat |
Embarrassingly Parallel Inference for Gaussian Processes | Training Gaussian process-based models typically involves an $ O(N^3)$
computational bottleneck due to inverting the covariance matrix. Popular
methods for overcoming this matrix inversion problem cannot adequately model
all types of latent functions, and are often not parallelizable. However,
judicious choice of model structure can ameliorate this problem. A
mixture-of-experts model that uses a mixture of $K$ Gaussian processes offers
modeling flexibility and opportunities for scalable inference. Our
embarrassingly parallel algorithm combines low-dimensional matrix inversions
with importance sampling to yield a flexible, scalable mixture-of-experts model
that offers comparable performance to Gaussian process regression at a much
lower computational cost.
| stat |
Computation for Latent Variable Model Estimation: A Unified Stochastic
Proximal Framework | Latent variable models have been playing a central role in psychometrics and
related fields. In many modern applications, the inference based on latent
variable models involves one or several of the following features: (1) the
presence of many latent variables, (2) the observed and latent variables being
continuous, discrete, or a combination of both, (3) constraints on parameters,
and (4) penalties on parameters to impose model parsimony. The estimation often
involves maximizing an objective function based on a marginal
likelihood/pseudo-likelihood, possibly with constraints and/or penalties on
parameters. Solving this optimization problem is highly non-trivial, due to the
complexities brought by the features mentioned above. Although several
efficient algorithms have been proposed, there lacks a unified computational
framework that takes all these features into account. In this paper, we fill
the gap. Specifically, we provide a unified formulation for the optimization
problem and then propose a quasi-Newton stochastic proximal algorithm.
Theoretical properties of the proposed algorithms are established. The
computational efficiency and robustness are shown by simulation studies under
various settings for latent variable model estimation.
| stat |
Large Volatility Matrix Prediction with High-Frequency Data | We provide a novel method for large volatility matrix prediction with
high-frequency data by applying eigen-decomposition to daily realized
volatility matrix estimators and capturing eigenvalue dynamics with ARMA
models. Given a sequence of daily volatility matrix estimators, we compute the
aggregated eigenvectors and obtain the corresponding eigenvalues. Eigenvalues
in the same relative magnitude form a time series and the ARMA models are
further employed to model the dynamics within each eigenvalue time series to
produce a predictor. We predict future large volatility matrix based on the
predicted eigenvalues and the aggregated eigenvectors, and demonstrate the
advantages of the proposed method in volatility prediction and portfolio
allocation problems.
| stat |
Rapid Robust Principal Component Analysis: CUR Accelerated Inexact Low
Rank Estimation | Robust principal component analysis (RPCA) is a widely used tool for
dimension reduction. In this work, we propose a novel non-convex algorithm,
coined Iterated Robust CUR (IRCUR), for solving RPCA problems, which
dramatically improves the computational efficiency in comparison with the
existing algorithms. IRCUR achieves this acceleration by employing CUR
decomposition when updating the low rank component, which allows us to obtain
an accurate low rank approximation via only three small submatrices.
Consequently, IRCUR is able to process only the small submatrices and avoid
expensive computing on the full matrix through the entire algorithm. Numerical
experiments establish the computational advantage of IRCUR over the
state-of-art algorithms on both synthetic and real-world datasets.
| stat |
Sequential adaptive strategy for population-based sampling of a rare and
clustered disease | An innovative sampling strategy is proposed, which applies to large-scale
population-based surveys targeting a rare trait that is unevenly spread over a
geographical area of interest. Our proposal is characterised by the ability to
tailor the data collection to specific features and challenges of the survey at
hand. It is based on integrating an adaptive component into a sequential
selection, which aims to both intensify detection of positive cases, upon
exploiting the spatial clusterisation, and provide a flexible framework for
managing logistical and budget constraints. To account for the selection bias,
a ready-to-implement weighting system is provided to release unbiased and
accurate estimates. Empirical evidence is illustrated from tuberculosis
prevalence surveys, which are recommended in many countries and supported by
the WHO as an emblematic example of the need for an improved sampling design.
Simulation results are also given to illustrate strengths and weaknesses of the
proposed sampling strategy with respect to traditional cross-sectional
sampling.
| stat |
The need for adequate sampling in a well-functioning market surveillance
system | Adequate sampling is essential for the well-functioning of a market
surveillance system. As small as possible statistically significant sample size
is the main factor that determines the costs of market surveillance actions.
This paper studies various possibilities for calculation of the size of the
sample with an emphasis on the method based on the binomial distribution.
Examples, comparisons, and conclusions are provided.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.