title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
A Comparison of Likelihood-Free Methods With and Without Summary
Statistics | Likelihood-free methods are useful for parameter estimation of complex models
with intractable likelihood functions for which it is easy to simulate data.
Such models are prevalent in many disciplines including genetics, biology,
ecology and cosmology. Likelihood-free methods avoid explicit likelihood
evaluation by finding parameter values of the model that generate data close to
the observed data. The general consensus has been that it is most efficient to
compare datasets on the basis of a low dimensional informative summary
statistic, incurring information loss in favour of reduced dimensionality. More
recently, researchers have explored various approaches for efficiently
comparing empirical distributions in the likelihood-free context in an effort
to avoid data summarisation. This article provides a review of these full data
distance based approaches, and conducts the first comprehensive comparison of
such methods, both qualitatively and empirically. We also conduct a substantive
empirical comparison with summary statistic based likelihood-free methods. The
discussion and results offer guidance to practitioners considering a
likelihood-free approach. Whilst we find the best approach to be problem
dependent, we also find that the full data distance based approaches are
promising and warrant further development. We discuss some opportunities for
future research in this space.
| stat |
Temporal Fusion Transformers for Interpretable Multi-horizon Time Series
Forecasting | Multi-horizon forecasting problems often contain a complex mix of inputs --
including static (i.e. time-invariant) covariates, known future inputs, and
other exogenous time series that are only observed historically -- without any
prior information on how they interact with the target. While several deep
learning models have been proposed for multi-step prediction, they typically
comprise black-box models which do not account for the full range of inputs
present in common scenarios. In this paper, we introduce the Temporal Fusion
Transformer (TFT) -- a novel attention-based architecture which combines
high-performance multi-horizon forecasting with interpretable insights into
temporal dynamics. To learn temporal relationships at different scales, the TFT
utilizes recurrent layers for local processing and interpretable self-attention
layers for learning long-term dependencies. The TFT also uses specialized
components for the judicious selection of relevant features and a series of
gating layers to suppress unnecessary components, enabling high performance in
a wide range of regimes. On a variety of real-world datasets, we demonstrate
significant performance improvements over existing benchmarks, and showcase
three practical interpretability use-cases of TFT.
| stat |
Post-processing multi-ensemble temperature and precipitation forecasts
through an Exchangeable Gamma Normal model and its Tobit extension | Meteorological ensembles are a collection of scenarios for future weather
delivered by a meteorological center. Such ensembles form the main source of
valuable information for probabilistic forecasting which aims at producing a
predictive probability distribution of the quantity of interest instead of a
single best guess estimate. Unfortunately, ensembles cannot generally be
considered as a sample from such a predictive probability distribution without
a preliminary post-processing treatment to calibrate the ensemble. Two main
families of post-processing methods, either competing such as BMA or
collaborative such as EMOS, can be found in the literature. This paper proposes
a mixed effect model belonging to the collaborative family. The structure of
the model is based on the hypothesis of invariance under the relabelling of the
ensemble members. Its interesting specificities are as follows: 1)
exchangeability, which contributes to parsimony, with a latent pivot variable
synthesizing the essential meteorological features of the ensembles, 2) a
multi-ensemble implementation, allowing to take advantage of various
information so as to increase the sharpness of the forecasting procedure. Focus
is cast onto Normal statistical structures, first with a direct application for
temperatures, then with its Tobit extension for precipitation. Inference is
performed by EM algorithms with recourse made to stochastic conditional
simulations in the precipitation case. After checking its good behavior on
artificial data, the proposed post-processing technique is applied to
temperature and precipitation ensemble forecasts produced over five river
basins managed by Hydro-Qu$\'e$bec. These ensemble forecasts were extracted
from the THORPEX Interactive Grand Global Ensemble (TIGGE) database. The
results indicate that post-processed ensemble are calibrated and generally
sharper than the raw ensembles.
| stat |
Simulating feedback mechanisms in patient flow and return visits in an
emergency department | Emergency department (ED) crowding has been an increasing problem worldwide.
Prior research has identified factors that contribute to ED crowding. However,
the relationships between these remain incompletely understood. This study's
objective was to analyse the effects of initiating a local protocol to
alleviate crowding situations at the expense of increasing returning patients
through the development of a system dynamics (SD) simulation model. The SD
study is from an academic care hospital in Boston, MA. Data sources include
direct observations, semi-structured interviews, archival data from October
2013, and peer-reviewed literature from the domains of emergency medicine and
management science. The SD model shows interrelations between inpatient
capacity restraints and return visits due to potential premature discharges.
The model reflects the vulnerability of the ED system when exposed to
unpredicted increases in demand. Default trigger values for the protocol are
tested to determine a balance between increased patient flows and the number of
returning patients. Baseline simulation runs for generic variables assessment
showed high leverage potential in bed assignment- and transfer times.
A thorough understanding of the complex non-linear behaviour of causes and
effects of ED crowding is enabled through the use of SD. The vulnerability of
the system lies in the crucial interaction between the physical constraints and
the expedited patient flows through protocol activation. This study is an
example of how hospital managers can benefit from virtual scenario testing
within a safe simulation environment to immediately visualise the impacts of
policy adjustments.
| stat |
A study of dependency features of spike trains through copulas | Simultaneous recordings from many neurons hide important information and the
connections characterizing the network remain generally undiscovered despite
the progresses of statistical and machine learning techniques. Discerning the
presence of direct links between neuron from data is still a not completely
solved problem. To enlarge the number of tools for detecting the underlying
network structure, we propose here the use of copulas, pursuing on a research
direction we started in [1]. Here, we adapt their use to distinguish different
types of connections on a very simple network. Our proposal consists in
choosing suitable random intervals in pairs of spike trains determining the
shapes of their copulas. We show that this approach allows to detect different
types of dependencies. We illustrate the features of the proposed method on
synthetic data from suitably connected networks of two or three formal neurons
directly connected or influenced by the surrounding network. We show how a
smart choice of pairs of random times together with the use of empirical
copulas allows to discern between direct and un-direct interactions.
| stat |
Variational Bayesian Methods for Stochastically Constrained System
Design Problems | We study system design problems stated as parameterized stochastic programs
with a chance-constraint set. We adopt a Bayesian approach that requires the
computation of a posterior predictive integral which is usually intractable. In
addition, for the problem to be a well-defined convex program, we must retain
the convexity of the feasible set. Consequently, we propose a variational
Bayes-based method to approximately compute the posterior predictive integral
that ensures tractability and retains the convexity of the feasible set. Under
certain regularity conditions, we also show that the solution set obtained
using variational Bayes converges to the true solution set as the number of
observations tends to infinity. We also provide bounds on the probability of
qualifying a true infeasible point (with respect to the true constraints) as
feasible under the VB approximation for a given number of samples.
| stat |
Spectral Density-Based and Measure-Preserving ABC for partially observed
diffusion processes. An illustration on Hamiltonian SDEs | Approximate Bayesian Computation (ABC) has become one of the major tools of
likelihood-free statistical inference in complex mathematical models.
Simultaneously, stochastic differential equations (SDEs) have developed to an
established tool for modelling time dependent, real world phenomena with
underlying random effects. When applying ABC to stochastic models, two major
difficulties arise. First, the derivation of effective summary statistics and
proper distances is particularly challenging, since simulations from the
stochastic process under the same parameter configuration result in different
trajectories. Second, exact simulation schemes to generate trajectories from
the stochastic model are rarely available, requiring the derivation of suitable
numerical methods for the synthetic data generation. To obtain summaries that
are less sensitive to the intrinsic stochasticity of the model, we propose to
build up the statistical method (e.g., the choice of the summary statistics) on
the underlying structural properties of the model. Here, we focus on the
existence of an invariant measure and we map the data to their estimated
invariant density and invariant spectral density. Then, to ensure that these
model properties are kept in the synthetic data generation, we adopt
measure-preserving numerical splitting schemes. The derived property-based and
measure-preserving ABC method is illustrated on the broad class of partially
observed Hamiltonian type SDEs, both with simulated data and with real
electroencephalography (EEG) data. The proposed ingredients can be incorporated
into any type of ABC algorithm and directly applied to all SDEs that are
characterised by an invariant distribution and for which a measure-preserving
numerical method can be derived.
| stat |
Dark Experience for General Continual Learning: a Strong, Simple
Baseline | Continual Learning has inspired a plethora of approaches and evaluation
settings; however, the majority of them overlooks the properties of a practical
scenario, where the data stream cannot be shaped as a sequence of tasks and
offline training is not viable. We work towards General Continual Learning
(GCL), where task boundaries blur and the domain and class distributions shift
either gradually or suddenly. We address it through mixing rehearsal with
knowledge distillation and regularization; our simple baseline, Dark Experience
Replay, matches the network's logits sampled throughout the optimization
trajectory, thus promoting consistency with its past. By conducting an
extensive analysis on both standard benchmarks and a novel GCL evaluation
setting (MNIST-360), we show that such a seemingly simple baseline outperforms
consolidated approaches and leverages limited resources. We further explore the
generalization capabilities of our objective, showing its regularization being
beneficial beyond mere performance.
| stat |
Bayesian Modeling of Microbiome Data for Differential Abundance Analysis | The advances of next-generation sequencing technology have accelerated study
of the microbiome and stimulated the high throughput profiling of metagenomes.
The large volume of sequenced data has encouraged the rise of various studies
for detecting differentially abundant taxonomic features across healthy and
diseased populations, with the ultimate goal of deciphering the relationship
between the microbiome diversity and health conditions. As the microbiome data
are high-dimensional, typically featuring by uneven sampling depth,
overdispersion and a huge amount of zeros, these data characteristics often
hamper the downstream analysis. Moreover, the taxonomic features are implicitly
imposed by the phylogenetic tree structure and often ignored. To overcome these
challenges, we propose a Bayesian hierarchical modeling framework for the
analysis of microbiome count data for differential abundance analysis. Under
this framework, we introduce a bi-level Bayesian hierarchical model that allows
a flexible choice of the count generating process, and hyperpriors in the
feature selection scheme. We particularly focus on employing a zero-inflated
negative binomial model with a Bayesian nonparametric prior model on the bottom
level, and applying Gaussian mixture models for differentially abundant taxa
detection on the top level. Our method allows for the simultaneous modeling of
sample heterogeneity and detecting differentially abundant taxa. We conducted
comprehensive simulations and summarized the improved statistical performances
of the proposed model. We applied the model in two real microbiome study
datasets and successfully identified biologically validated differentially
abundant taxa. We hope that the proposed framework and model can facilitate
further microbiome studies and elucidate disease etiology.
| stat |
Distribution-on-Distribution Regression via Optimal Transport Maps | We present a framework for performing regression when both covariate and
response are probability distributions on a compact interval
$\Omega\subset\mathbb{R}$. Our regression model is based on the theory of
optimal transportation and links the conditional Fr\'echet mean of the response
distribution to the covariate distribution via an optimal transport map. We
define a Fr\'echet-least-squares estimator of this regression map, and
establish its consistency and rate of convergence to the true map, under both
full and partial observation of the regression pairs. Computation of the
estimator is shown to reduce to an isotonic regression problem, and thus our
regression model can be implemented with ease. We illustrate our methodology
using real and simulated data.
| stat |
Inference in generalized bilinear models | Latent factor models are widely used to discover and adjust for hidden
variation in modern applications. However, most methods do not fully account
for uncertainty in the latent factors, which can lead to miscalibrated
inferences such as overconfident p-values. In this article, we develop a fast
and accurate method of uncertainty quantification in generalized bilinear
models, which are a flexible extension of generalized linear models to include
latent factors as well as row covariates, column covariates, and interactions.
In particular, we introduce delta propagation, a general technique for
propagating uncertainty among model components using the delta method. Further,
we provide a rapidly converging algorithm for maximum a posteriori GBM
estimation that extends earlier methods by estimating row and column
dispersions. In simulation studies, we find that our method provides
approximately correct frequentist coverage of most parameters of interest. We
demonstrate on RNA-seq gene expression analysis and copy ratio estimation in
cancer genomics.
| stat |
A Generalization of the Savage-Dickey Density Ratio for Testing Equality
and Order Constrained Hypotheses | The Savage-Dickey density ratio is a specific expression of the Bayes factor
when testing a precise (equality constrained) hypothesis against an
unrestricted alternative. The expression greatly simplifies the computation of
the Bayes factor at the cost of assuming a specific form of the prior under the
precise hypothesis as a function of the unrestricted prior. A generalization
was proposed by Verdinelli and Wasserman (1995) such that the priors can be
freely specified under both hypotheses while keeping the computational
advantage. This paper presents an extension of this generalization when the
hypothesis has equality as well as order constraints on the parameters of
interest. The methodology is used for a constrained multivariate t test using
the JZS Bayes factor and a constrained hypothesis test under the multinomial
model.
| stat |
Bayesian models for survival data of clinical trials: Comparison of
implementations using R software | Objective: To provide guidance for the use of the main functions available in
R for performing post hoc Bayesian analysis of a randomized clinical trial with
a survival endpoint using proportional hazard models. Study Design and Setting:
Data derived from the ALLOZITHRO trial, conducted with 465 patients after
allograft to prevent pulmonary complications and allocated between azithromycin
and placebo; airflow decline-free survival at 2 years after randomization was
the main endpoint. Results: Despite heterogeneity in modeling assumptions, in
particular for the baseline hazard (parametric or nonparametric), and in
estimation methods, Bayesian posterior mean hazard ratio (HR) estimates of
azithromycin effect were close to those obtained by the maximum likelihood
approach. Conclusion: Bayesian models can be implemented using various R
packages, providing results in close agreement with the maximum likelihood
estimates. These models provide probabilistic statements that could not be
obtained otherwise.
| stat |
Structural modeling using overlapped group penalties for discovering
predictive biomarkers for subgroup analysis | The identification of predictive biomarkers from a large scale of covariates
for subgroup analysis has attracted fundamental attention in medical research.
In this article, we propose a generalized penalized regression method with a
novel penalty function, for enforcing the hierarchy structure between the
prognostic and predictive effects, such that a nonzero predictive effect must
induce its ancestor prognostic effects being nonzero in the model. Our method
is able to select useful predictive biomarkers by yielding a sparse,
interpretable, and predictable model for subgroup analysis, and can deal with
different types of response variable such as continuous, categorical, and
time-to-event data. We show that our method is asymptotically consistent under
some regularized conditions. To minimize the generalized penalized regression
model, we propose a novel integrative optimization algorithm by integrating the
majorization-minimization and the alternating direction method of multipliers,
which is named after \texttt{smog}. The enriched simulation study and real case
study demonstrate that our method is very powerful for discovering the true
predictive biomarkers and identifying subgroups of patients.
| stat |
Exact Tests for Offline Changepoint Detection in Multichannel Binary and
Count Data with Application to Networks | We consider offline detection of a single changepoint in binary and count
time-series. We compare exact tests based on the cumulative sum (CUSUM) and the
likelihood ratio (LR) statistics, and a new proposal that combines exact
two-sample conditional tests with multiplicity correction, against standard
asymptotic tests based on the Brownian bridge approximation to the CUSUM
statistic. We see empirically that the exact tests are much more powerful in
situations where normal approximations driving asymptotic tests are not
trustworthy: (i) small sample settings; (ii) sparse parametric settings; (iii)
time-series with changepoint near the boundary.
We also consider a multichannel version of the problem, where channels can
have different changepoints. Controlling the False Discovery Rate (FDR), we
simultaneously detect changes in multiple channels. This "local" approach is
shown to be more advantageous than multivariate global testing approaches when
the number of channels with changepoints is much smaller than the total number
of channels.
As a natural application, we consider network-valued time-series and use our
approach with (a) edges as binary channels and (b) node-degrees or other local
subgraph statistics as count channels. The local testing approach is seen to be
much more informative than global network changepoint algorithms.
| stat |
Is Q-Learning Minimax Optimal? A Tight Sample Complexity Analysis | Q-learning, which seeks to learn the optimal Q-function of a Markov decision
process (MDP) in a model-free fashion, lies at the heart of reinforcement
learning. When it comes to the synchronous setting (such that independent
samples for all state-action pairs are drawn from a generative model in each
iteration), substantial progress has been made recently towards understanding
the sample efficiency of Q-learning. Take a $\gamma$-discounted
infinite-horizon MDP with state space $\mathcal{S}$ and action space
$\mathcal{A}$: to yield an entrywise $\varepsilon$-accurate estimate of the
optimal Q-function, state-of-the-art theory for Q-learning proves that a sample
size on the order of
$\frac{|\mathcal{S}||\mathcal{A}|}{(1-\gamma)^5\varepsilon^{2}}$ is sufficient,
which, however, fails to match with the existing minimax lower bound. This
gives rise to natural questions: what is the sharp sample complexity of
Q-learning? Is Q-learning provably sub-optimal? In this work, we settle these
questions by (1) demonstrating that the sample complexity of Q-learning is at
most on the order of
$\frac{|\mathcal{S}||\mathcal{A}|}{(1-\gamma)^4\varepsilon^2}$ (up to some log
factor) for any $0<\varepsilon <1$, and (2) developing a matching lower bound
to confirm the sharpness of our result. Our findings unveil both the
effectiveness and limitation of Q-learning: its sample complexity matches that
of speedy Q-learning without requiring extra computation and storage, albeit
still being considerably higher than the minimax lower bound.
| stat |
Mutual Information Gradient Estimation for Representation Learning | Mutual Information (MI) plays an important role in representation learning.
However, MI is unfortunately intractable in continuous and high-dimensional
settings. Recent advances establish tractable and scalable MI estimators to
discover useful representation. However, most of the existing methods are not
capable of providing an accurate estimation of MI with low-variance when the MI
is large. We argue that directly estimating the gradients of MI is more
appealing for representation learning than estimating MI in itself. To this
end, we propose the Mutual Information Gradient Estimator (MIGE) for
representation learning based on the score estimation of implicit
distributions. MIGE exhibits a tight and smooth gradient estimation of MI in
the high-dimensional and large-MI settings. We expand the applications of MIGE
in both unsupervised learning of deep representations based on InfoMax and the
Information Bottleneck method. Experimental results have indicated significant
performance improvement in learning useful representation.
| stat |
The Power of Unbiased Recursive Partitioning: A Unifying View of CTree,
MOB, and GUIDE | A core step of every algorithm for learning regression trees is the selection
of the best splitting variable from the available covariates and the
corresponding split point. Early tree algorithms (e.g., AID, CART) employed
greedy search strategies, directly comparing all possible split points in all
available covariates. However, subsequent research showed that this is biased
towards selecting covariates with more potential split points. Therefore,
unbiased recursive partitioning algorithms have been suggested (e.g., QUEST,
GUIDE, CTree, MOB) that first select the covariate based on statistical
inference using p-values that are adjusted for the possible split points. In a
second step a split point optimizing some objective function is selected in the
chosen split variable. However, different unbiased tree algorithms obtain these
p-values from different inference frameworks and their relative advantages or
disadvantages are not well understood, yet. Therefore, three different popular
approaches are considered here: classical categorical association tests (as in
GUIDE), conditional inference (as in CTree), and parameter instability tests
(as in MOB). First, these are embedded into a common inference framework
encompassing parametric model trees, in particular linear model trees. Second,
it is assessed how different building blocks from this common framework affect
the power of the algorithms to select the appropriate covariates for splitting:
observation-wise goodness-of-fit measure (residuals vs. model scores),
dichotomization of residuals/scores at zero, and binning of possible split
variables. This shows that specifically the goodness-of-fit measure is crucial
for the power of the procedures, with model scores without dichotomization
performing much better in many scenarios.
| stat |
Learning from Similarity-Confidence Data | Weakly supervised learning has drawn considerable attention recently to
reduce the expensive time and labor consumption of labeling massive data. In
this paper, we investigate a novel weakly supervised learning problem of
learning from similarity-confidence (Sconf) data, where we aim to learn an
effective binary classifier from only unlabeled data pairs equipped with
confidence that illustrates their degree of similarity (two examples are
similar if they belong to the same class). To solve this problem, we propose an
unbiased estimator of the classification risk that can be calculated from only
Sconf data and show that the estimation error bound achieves the optimal
convergence rate. To alleviate potential overfitting when flexible models are
used, we further employ a risk correction scheme on the proposed risk
estimator. Experimental results demonstrate the effectiveness of the proposed
methods.
| stat |
Segmenting High-dimensional Matrix-valued Time Series via Sequential
Transformations | Modeling matrix-valued time series is an interesting and important research
topic. In this paper, we extend the method of Chang et al. (2017) to
matrix-valued time series. For any given $p\times q$ matrix-valued time series,
we look for linear transformations to segment the matrix into many small
sub-matrices for which each of them are uncorrelated with the others both
contemporaneously and serially, thus they can be analyzed separately, which
will greatly reduce the number of parameters to be estimated in terms of
modeling. To overcome the identification issue, we propose a two-step and more
structured procedure to segment the rows and columns separately. When
$\max(p,q)$ is large in relation to the sample size $n$, we assume the
transformation matrices are sparse and use threshold estimators for the
(auto)covariance matrices. We also propose a block-wisely thresholding method
to separate the columns (or rows) of the transformed matrix-valued data. The
asymptotic properties are established for both fixed and diverging $\max(p,q)$.
Unlike principal component analysis (PCA) for independent data, we cannot
guarantee that the required linear transformation exists. When it does not, the
proposed method provides an approximate segmentation, which may be useful for
forecasting. The proposed method is illustrated with both simulated and real
data examples. We also propose a sequential transformation algorithm to segment
higher-order tensor-valued time series.
| stat |
Forecasting age distribution of death counts: An application to annuity
pricing | We consider a compositional data analysis approach to forecasting the age
distribution of death counts. Using the age-specific period life-table death
counts in Australia obtained from the Human Mortality Database, the
compositional data analysis approach produces more accurate one- to
20-step-ahead point and interval forecasts than Lee-Carter method,
Hyndman-Ullah method, and two na\"{i}ve random walk methods. The improved
forecast accuracy of period life-table death counts is of great interest to
demographers for estimating survival probabilities and life expectancy, and to
actuaries for determining temporary annuity prices for various ages and
maturities. Although we focus on temporary annuity prices, we consider
long-term contracts which make the annuity almost lifetime, in particular when
the age at entry is sufficiently high.
| stat |
A contribution to Optimal Transport on incomparable spaces | Optimal Transport is a theory that allows to define geometrical notions of
distance between probability distributions and to find correspondences,
relationships, between sets of points. Many machine learning applications are
derived from this theory, at the frontier between mathematics and optimization.
This thesis proposes to study the complex scenario in which the different data
belong to incomparable spaces. In particular we address the following
questions: how to define and apply Optimal Transport between graphs, between
structured data? How can it be adapted when the data are varied and not
embedded in the same metric space? This thesis proposes a set of Optimal
Transport tools for these different cases. An important part is notably devoted
to the study of the Gromov-Wasserstein distance whose properties allow to
define interesting transport problems on incomparable spaces. More broadly, we
analyze the mathematical properties of the various proposed tools, we establish
algorithmic solutions to compute them and we study their applicability in
numerous machine learning scenarii which cover, in particular, classification,
simplification, partitioning of structured data, as well as heterogeneous
domain adaptation.
| stat |
Practical and sample efficient zero-shot HPO | Zero-shot hyperparameter optimization (HPO) is a simple yet effective use of
transfer learning for constructing a small list of hyperparameter (HP)
configurations that complement each other. That is to say, for any given
dataset, at least one of them is expected to perform well. Current techniques
for obtaining this list are computationally expensive as they rely on running
training jobs on a diverse collection of datasets and a large collection of
randomly drawn HPs. This cost is especially problematic in environments where
the space of HPs is regularly changing due to new algorithm versions, or
changing architectures of deep networks. We provide an overview of available
approaches and introduce two novel techniques to handle the problem. The first
is based on a surrogate model and adaptively chooses pairs of dataset,
configuration to query. The second, for settings where finding, tuning and
testing a surrogate model is problematic, is a multi-fidelity technique
combining HyperBand with submodular optimization. We benchmark our methods
experimentally on five tasks (XGBoost, LightGBM, CatBoost, MLP and AutoML) and
show significant improvement in accuracy compared to standard zero-shot HPO
with the same training budget. In addition to contributing new algorithms, we
provide an extensive study of the zero-shot HPO technique resulting in (1)
default hyper-parameters for popular algorithms that would benefit the
community using them, (2) massive lookup tables to further the research of
hyper-parameter tuning.
| stat |
Helmholtzian Eigenmap: Topological feature discovery & edge flow
learning from point cloud data | The manifold Helmholtzian (1-Laplacian) operator $\Delta_1$ elegantly
generalizes the Laplace-Beltrami operator to vector fields on a manifold
$\mathcal M$. In this work, we propose the estimation of the manifold
Helmholtzian from point cloud data by a weighted 1-Laplacian $\mathbf{\mathcal
L}_1$. While higher order Laplacians ave been introduced and studied, this work
is the first to present a graph Helmholtzian constructed from a simplicial
complex as an estimator for the continuous operator in a non-parametric
setting. Equipped with the geometric and topological information about
$\mathcal M$, the Helmholtzian is a useful tool for the analysis of flows and
vector fields on $\mathcal M$ via the Helmholtz-Hodge theorem. In addition, the
$\mathbf{\mathcal L}_1$ allows the smoothing, prediction, and feature
extraction of the flows. We demonstrate these possibilities on substantial sets
of synthetic and real point cloud datasets with non-trivial topological
structures; and provide theoretical results on the limit of $\mathbf{\mathcal
L}_1$ to $\Delta_1$.
| stat |
Learning Proposals for Probabilistic Programs with Inference Combinators | We develop operators for construction of proposals in probabilistic programs,
which we refer to as inference combinators. Inference combinators define a
grammar over importance samplers that compose primitive operations such as
application of a transition kernel and importance resampling. Proposals in
these samplers can be parameterized using neural networks, which in turn can be
trained by optimizing variational objectives. The result is a framework for
user-programmable variational methods that are correct by construction and can
be tailored to specific models. We demonstrate the flexibility of this
framework by implementing advanced variational methods based on amortized Gibbs
sampling and annealing.
| stat |
Beta Rank Function: A Smooth Double-Pareto-Like Distribution | The Beta Rank Function (BRF) $x(u) =A(1-u)^b/u^a$, where $u$ is the
normalized and continuous rank of an observation $x$, has wide applications in
fitting real-world data from social science to biological phenomena. The
underlying probability density function (pdf) $f_X(x)$ does not usually have a
closed expression except for specific parameter values. We show however that it
is approximately a unimodal skewed and asymmetric two-sided power law/double
Pareto/log-Laplacian distribution. The BRF pdf has simple properties when the
independent variable is log-transformed: $f_{Z=\log(X)}(z)$ . At the peak it
makes a smooth turn and it does not diverge, lacking the sharp angle observed
in the double Pareto or Laplace distribution. The peak position of $f_Z(z)$ is
$z_0=\log A+(a-b)\log(\sqrt{a}+\sqrt{b})-(a\log(a)-b\log(b))/2 $; the
probability is partitioned by the peak to the proportion of
$\sqrt{b}/(\sqrt{a}+\sqrt{b})$ (left) and $\sqrt{a}/(\sqrt{a}+\sqrt{b})$
(right); the functional form near the peak is controlled by the cubic term in
the Taylor expansion when $a\ne b$; the mean of $Z$ is $E[Z]=\log A+a-b$; the
decay on left and right sides of the peak is approximately exponential with
forms $e^{\frac{z-\log A}{b} }/b$ and $e^{ -\frac{z-\log A}{a}}/a$. These
results are confirmed by numerical simulations. Properties of $f_X(x)$ without
log-transforming the variable are much more complex, though the approximate
double Pareto behavior, $(x/A)^{1/b}/(bx)$ (for $x<A$) and $(x/A)^{-1/a}/(ax)$
(for $x > A$) is simple. Our results elucidate the relationship between BRF and
log-normal distributions when $a=b$ and explain why the BRF is ubiquitous and
versatile. Based on the pdf, we suggest a quick way to elucidate if a real data
set follows a one-sided power-law, a log-normal, a two-sided power-law or a
BRF. We illustrate our results with two examples: urban populations and
financial returns.
| stat |
Prune Sampling: a MCMC inference technique for discrete and
deterministic Bayesian networks | We introduce and characterise the performance of the Markov chain Monte Carlo
(MCMC) inference method Prune Sampling for discrete and deterministic Bayesian
networks (BNs). We developed a procedure to obtain the performance of a MCMC
sampling method in the limit of infinite simulation time, extrapolated from
relatively short simulations. This approach was used to conduct a study to
compare the accuracy, rate of convergence and the time consumption of Prune
Sampling with two conventional MCMC sampling methods: Gibbs- and Metropolis
sampling. We show that Markov chains created by Prune Sampling always converge
to the desired posterior distribution, also for networks where conventional
Gibbs sampling fails. Beside this, we demonstrate that pruning outperforms
Gibbs sampling, at least for a certain class of BNs. Though, this tempting
feature comes at a price. In the first version of Prune Sampling, for large BNs
the procedure to choose the next iteration step uniformly is rather time
intensive. Our conclusion is that Prune Sampling is a competitive method for
all types of small and medium sized BNs, but (for now) standard methods still
perform better for all types of large BNs.
| stat |
Automatic structured variational inference | Stochastic variational inference offers an attractive option as a default
method for differentiable probabilistic programming. However, the performance
of the variational approach depends on the choice of an appropriate variational
family. Here, we introduce automatic structured variational inference (ASVI), a
fully automated method for constructing structured variational families,
inspired by the closed-form update in conjugate Bayesian models. These
convex-update families incorporate the forward pass of the input probabilistic
program and can therefore capture complex statistical dependencies.
Convex-update families have the same space and time complexity as the input
probabilistic program and are therefore tractable for a very large family of
models including both continuous and discrete variables. We validate our
automatic variational method on a wide range of low- and high-dimensional
inference problems. We find that ASVI provides a clear improvement in
performance when compared with other popular approaches such as the mean-field
approach and inverse autoregressive flows. We provide an open source
implementation of ASVI in TensorFlow Probability.
| stat |
A computational method for estimating Burr XII parameters with complete
and multiple censored data | Flexibility in shape and scale of Burr XII distribution can make close
approximation of numerous well-known probability density functions. Due to
these capabilities, the usages of Burr XII distribution are applied in risk
analysis, lifetime data analysis and process capability estimation. In this
paper the Cross-Entropy (CE) method is further developed in terms of Maximum
Likelihood Estimation (MLE) to estimate the parameters of Burr XII distribution
for the complete data or in the presence of multiple censoring. A simulation
study is conducted to evaluate the performance of the MLE by means of CE method
for different parameter settings and sample sizes. The results are compared to
other existing methods in both uncensored and censored situations.
| stat |
Interpretable Principal Components Analysis for Multilevel Multivariate
Functional Data, with Application to EEG Experiments | Many studies collect functional data from multiple subjects that have both
multilevel and multivariate structures. An example of such data comes from
popular neuroscience experiments where participants' brain activity is recorded
using modalities such as EEG and summarized as power within multiple
time-varying frequency bands within multiple electrodes, or brain regions.
Summarizing the joint variation across multiple frequency bands for both
whole-brain variability between subjects, as well as location-variation within
subjects, can help to explain neural reactions to stimuli. This article
introduces a novel approach to conducting interpretable principal components
analysis on multilevel multivariate functional data that decomposes total
variation into subject-level and replicate-within-subject-level (i.e.
electrode-level) variation, and provides interpretable components that can be
both sparse among variates (e.g. frequency bands) and have localized support
over time within each frequency band. The sparsity and localization of
components is achieved by solving an innovative rank-one based convex
optimization problem with block Frobenius and matrix $L_1$-norm based
penalties. The method is used to analyze data from a study to better understand
reactions to emotional information in individuals with histories of trauma and
the symptom of dissociation, revealing new neurophysiological insights into how
subject- and electrode-level brain activity are associated with these
phenomena.
| stat |
Learning Dependency Structures for Weak Supervision Models | Labeling training data is a key bottleneck in the modern machine learning
pipeline. Recent weak supervision approaches combine labels from multiple noisy
sources by estimating their accuracies without access to ground truth labels;
however, estimating the dependencies among these sources is a critical
challenge. We focus on a robust PCA-based algorithm for learning these
dependency structures, establish improved theoretical recovery rates, and
outperform existing methods on various real-world tasks. Under certain
conditions, we show that the amount of unlabeled data needed can scale
sublinearly or even logarithmically with the number of sources $m$, improving
over previous efforts that ignore the sparsity pattern in the dependency
structure and scale linearly in $m$. We provide an information-theoretic lower
bound on the minimum sample complexity of the weak supervision setting. Our
method outperforms weak supervision approaches that assume
conditionally-independent sources by up to 4.64 F1 points and previous
structure learning approaches by up to 4.41 F1 points on real-world relation
extraction and image classification tasks.
| stat |
Cross validation approaches for penalized Cox regression | Cross validation is commonly used for selecting tuning parameters in
penalized regression, but its use in penalized Cox regression models has
received relatively little attention in the literature. Due to its partial
likelihood construction, carrying out cross validation for Cox models is not
straightforward, and there are several potential approaches for implementation.
Here, we propose two new cross-validation methods for Cox regression and
compare them to approaches that have been proposed elsewhere. Our proposed
approach of cross-validating the linear predictors seems to offer an attractive
balance of performance and numerical stability. We illustrate these advantages
using simulated data as well as using them to analyze data from a
high-dimensional study of survival in lung cancer patients.
| stat |
Unifying Gaussian LWF and AMP Chain Graphs to Model Interference | An intervention may have an effect on units other than those to which it was
administered. This phenomenon is called interference and it usually goes
unmodeled. In this paper, we propose to combine Lauritzen-Wermuth-Frydenberg
and Andersson-Madigan-Perlman chain graphs to create a new class of causal
models that can represent both interference and non-interference relationships
for Gaussian distributions. Specifically, we define the new class of models,
introduce global and local and pairwise Markov properties for them, and prove
their equivalence. We also propose an algorithm for maximum likelihood
parameter estimation for the new models, and report experimental results.
Finally, we show how to compute the effects of interventions in the new models.
| stat |
Markov-chain Monte-Carlo Sampling for Optimal Fidelity Determination in
Dynamic Decision-Making | Decision making for dynamic systems is challenging due to the scale and
dynamicity of such systems, and it is comprised of decisions at strategic,
tactical, and operational levels. One of the most important aspects of decision
making is incorporating real time information that reflects immediate status of
the system. This type of decision making, which may apply to any dynamic
system, needs to comply with the system's current capabilities and calls for a
dynamic data driven planning framework. Performance of dynamic data driven
planning frameworks relies on the decision making process which in return is
relevant to the quality of the available data. This means that the planning
framework should be able to set the level of decision making based on the
current status of the system, which is learned through the continuous readings
of sensory data. In this work, a Markov chain Monte Carlo sampling method is
proposed to determine the optimal fidelity of decision making in a dynamic data
driven framework. To evaluate the performance of the proposed method, an
experiment is conducted, where the impact of workers performance on the
production capacity and the fidelity level of decision making are studied.
| stat |
$K$-Means and Gaussian Mixture Modeling with a Separation Constraint | We consider the problem of clustering with $K$-means and Gaussian mixture
models with a constraint on the separation between the centers in the context
of real-valued data. We first propose a dynamic programming approach to solving
the $K$-means problem with a separation constraint on the centers, building on
(Wang and Song, 2011). In the context of fitting a Gaussian mixture model, we
then propose an EM algorithm that incorporates such a constraint. A separation
constraint can help regularize the output of a clustering algorithm, and we
provide both simulated and real data examples to illustrate this point.
| stat |
Exploratory and Confirmatory Factor Analyses of Religiosity. A
Four-Factor Conceptual Model | We describe an exploratory and confirmatory factor analysis of the
International Social Survey Programme Religion Cumulation (1991-1998-2008) data
set, to identify the factors of individual religiosity and their interrelations
in quantitative terms. The exploratory factor analysis was performed using data
from the first two waves (1991 and 1998), and led to the identification of four
strongly correlated and reliable factors which we labeled Religious formation,
Supernatural beliefs, Belief in God, and Religious practice. The confirmatory
factor analysis was run using data from 2008, and led to the confirmation of
this four-factor structure with very good fit measures. We also ran a set of
structural equation models in an attempt to determine the causality links
between these four factors. It was found that for the models which provide the
best fit Belief in God does not cause Supernatural beliefs, Religious practice
can cause Belief in God and that there are multiple paths leading to Belief in
God, most of which include Religious formation as a source. The exploratory
factor analysis also led to the identification of other factors related to
traditional values, confidence in institutions and influence of religious
leaders on politics, but these were found to have lower reliability or
insufficient number of items to meet the acceptance criteria, and thus were not
included in the confirmatory factor analysis and the investigation of causal
links. The results obtained in this work have important material implications
for the conceptualization of "religiosity," and important methodological
implications for the scientific study of religion.
| stat |
Orthogonal Nonnegative Tucker Decomposition | In this paper, we study the nonnegative tensor data and propose an orthogonal
nonnegative Tucker decomposition (ONTD). We discuss some properties of ONTD and
develop a convex relaxation algorithm of the augmented Lagrangian function to
solve the optimization problem. The convergence of the algorithm is given. We
employ ONTD on the image data sets from the real world applications including
face recognition, image representation, hyperspectral unmixing. Numerical
results are shown to illustrate the effectiveness of the proposed algorithm.
| stat |
Constructing a Chain Event Graph from a Staged Tree | Chain Event Graphs (CEGs) are a recent family of probabilistic graphical
models - a generalisation of Bayesian Networks - providing an explicit
representation of structural zeros and context-specific conditional
independences within their graph topology. A CEG is constructed from an event
tree through a sequence of transformations beginning with the colouring of the
vertices of the event tree to identify one-step transition symmetries. This
coloured event tree, also known as a staged tree, is the output of the learning
algorithms used for this family. Surprisingly, no general algorithm has yet
been devised that automatically transforms any staged tree into a CEG
representation. In this paper we provide a simple iterative backward algorithm
for this transformation. Additionally, we show that no information is lost from
transforming a staged tree into a CEG. Finally, we demonstrate that with an
optimal stopping time, our algorithm is more efficient than the generalisation
of a special case presented in Silander and Leong (2013). We also provide
Python code using this algorithm to obtain a CEG from any staged tree along
with the functionality to add edges with sampling zeros.
| stat |
Probabilistic Modeling of Hurricane Wind-Induced Damage in
Infrastructure Systems | This paper presents a modeling approach for probabilistic estimation of
hurricane wind-induced damage to infrastructural assets. In our approach, we
employ a Nonhomogeneous Poisson Process (NHPP) model for estimating
spatially-varying probability distributions of damage as a function of
hurricane wind field velocities. Specifically, we consider a physically-based,
quadratic NHPP model for failures of overhead assets in electricity
distribution systems. The wind field velocities are provided by Forecasts of
Hurricanes using Large-Ensemble Outputs (FHLO), a framework for generating
probabilistic hurricane forecasts. We use FHLO in conjunction with the NHPP
model, such that the hurricane forecast uncertainties represented by FHLO are
accounted for in estimating the probability distributions of damage.
Furthermore, we evaluate the spatial variability and extent of hurricane damage
under key wind field parameters (intensity, size, and asymmetries). By applying
our approach to prediction of power outages (loss-of-service) in northwestern
Florida due to Hurricane Michael (2018), we demonstrate a statistically
significant relationship between outage rate and failure rate. Finally, we
formulate parametric models that relate total damage and financial losses to
the hurricane parameters of intensity and size. Overall, this paper's findings
suggest that our approach is well-suited to jointly account for spatial
variability and forecast uncertainty in the damage estimates, and is readily
applicable to prediction of system loss-of-service due to the damage.
| stat |
Block Length Choice for the Bootstrap of Dependent Panel Data -- a
Comment on Choi and Shin (2020) | Choi and Shin (2020) have constructed a bootstrap-based test for
change-points in panels with temporal and and/or cross-sectional dependence.
They have compared their test to several other proposed tests. We demonstrate
that by an appropriate, data-adaptive choice of the block length, the
change-point test by Sharipov, Tewes, Wendler (2016) can at least cope with
mild temporal dependence, the size distortion of this test is not as severe as
claimed by Choi and Shin (2020).
| stat |
Communication-Efficient Distributed SVD via Local Power Iterations | We study distributed computing of the truncated singular value decomposition
problem. We develop an algorithm that we call \texttt{LocalPower} for improving
communication efficiency. Specifically, we uniformly partition the dataset
among $m$ nodes and alternate between multiple (precisely $p$) local power
iterations and one global aggregation. In the aggregation, we propose to weight
each local eigenvector matrix with orthogonal Procrustes transformation (OPT).
As a practical surrogate of OPT, sign-fixing, which uses a diagonal matrix with
$\pm 1$ entries as weights, has better computation complexity and stability. We
theoretically show that under certain assumptions \texttt{LocalPower} lowers
the required number of communications by a factor of $p$ to reach a constant
accuracy. We also show that the strategy of periodically decaying $p$ helps
obtain high-precision solutions. We conduct experiments to demonstrate the
effectiveness of \texttt{LocalPower}.
| stat |
Wavelet Spatio-Temporal Change Detection on multi-temporal PolSAR images | We introduce WECS (Wavelet Energies Correlation Sreening), an unsupervised
sparse procedure to detect spatio-temporal change points on multi-temporal SAR
(POLSAR) images or even on sequences of very high resolution images. The
procedure is based on wavelet approximation for the multi-temporal images,
wavelet energy apportionment, and ultra-high dimensional correlation screening
for the wavelet coefficients. We present two complimentary wavelet measures in
order to detect sudden and/or cumulative changes, as well as for the case of
stationary or non-stationary multi-temporal images. We show WECS performance on
synthetic multi-temporal image data. We also apply the proposed method to a
time series of 85 satellite images in the border region of Brazil and the
French Guiana. The images were captured from November 08, 2015 to December 09
2017.
| stat |
Comparing Samples from the $\mathcal{G}^0$ Distribution using a Geodesic
Distance | The $\mathcal{G}^0$ distribution is widely used for monopolarized SAR image
modeling because it can characterize regions with different degree of texture
accurately. It is indexed by three parameters: the number of looks (which can
be estimated for the whole image), a scale parameter and a texture parameter.
This paper presents a new proposal for comparing samples from the
$\mathcal{G}^0$ distribution using a Geodesic Distance (GD) as a measure of
dissimilarity between models. The objective is quantifying the difference
between pairs of samples from SAR data using both local parameters (scale and
texture) of the $\mathcal{G}^0$ distribution. We propose three tests based on
the GD which combine the tests presented in~\cite{GeodesicDistanceGI0JSTARS},
and we estimate their probability distributions using permutation methods.
| stat |
On the Beta Prime Prior for Scale Parameters in High-Dimensional
Bayesian Regression Models | We study high-dimensional Bayesian linear regression with a general beta
prime distribution for the scale parameter. Under the assumption of sparsity,
we show that appropriate selection of the hyperparameters in the beta prime
prior leads to the (near) minimax posterior contraction rate when $p \gg n$.
For finite samples, we propose a data-adaptive method for estimating the
hyperparameters based on marginal maximum likelihood (MML). This enables our
prior to adapt to both sparse and dense settings, and under our proposed
empirical Bayes procedure, the MML estimates are never at risk of collapsing to
zero. We derive efficient Monte Carlo EM and variational EM algorithms for
implementing our model, which are available in the R package NormalBetaPrime.
Simulations and analysis of a gene expression data set illustrate our model's
self-adaptivity to varying levels of sparsity and signal strengths.
| stat |
Detection of similar successive groups in a model with diverging number
of variable groups | In this paper, a linear model with grouped explanatory variables is
considered. The idea is to perform an automatic detection of different
successive groups of the unknown coefficients under the assumption that the
number of groups is of the same order as the sample size. The standard least
squares loss function and the quantile loss function are both used together
with the fused and adaptive fused penalty to simultaneously estimate and group
the unknown parameters. The proper convergence rate is given for the obtained
estimators and the upper bound for the number of different successive group is
derived. A simulation study is used to compare the empirical performance of the
proposed fused and adaptive fused estimators and a real application on the air
quality data demonstrates the practical applicability of the proposed methods.
| stat |
Online Binary Space Partitioning Forests | The Binary Space Partitioning-Tree~(BSP-Tree) process was recently proposed
as an efficient strategy for space partitioning tasks. Because it uses more
than one dimension to partition the space, the BSP-Tree Process is more
efficient and flexible than conventional axis-aligned cutting strategies.
However, due to its batch learning setting, it is not well suited to
large-scale classification and regression problems. In this paper, we develop
an online BSP-Forest framework to address this limitation. With the arrival of
new data, the resulting online algorithm can simultaneously expand the space
coverage and refine the partition structure, with guaranteed universal
consistency for both classification and regression problems. The effectiveness
and competitive performance of the online BSP-Forest is verified via
simulations on real-world datasets.
| stat |
Semiparametric time series models driven by latent factor | We introduce a class of semiparametric time series models by assuming a
quasi-likelihood approach driven by a latent factor process. More specifically,
given the latent process, we only specify the conditional mean and variance of
the time series and enjoy a quasi-likelihood function for estimating parameters
related to the mean. This proposed methodology has three remarkable features:
(i) no parametric form is assumed for the conditional distribution of the time
series given the latent process; (ii) able for modelling non-negative, count,
bounded/binary and real-valued time series; (iii) dispersion parameter is not
assumed to be known. Further, we obtain explicit expressions for the marginal
moments and for the autocorrelation function of the time series process so that
a method of moments can be employed for estimating the dispersion parameter and
also parameters related to the latent process. Simulated results aiming to
check the proposed estimation procedure are presented. Real data analysis on
unemployment rate and precipitation time series illustrate the potencial for
practice of our methodology.
| stat |
Learning Multiple Defaults for Machine Learning Algorithms | The performance of modern machine learning methods highly depends on their
hyperparameter configurations. One simple way of selecting a configuration is
to use default settings, often proposed along with the publication and
implementation of a new algorithm. Those default values are usually chosen in
an ad-hoc manner to work good enough on a wide variety of datasets. To address
this problem, different automatic hyperparameter configuration algorithms have
been proposed, which select an optimal configuration per dataset. This
principled approach usually improves performance but adds additional
algorithmic complexity and computational costs to the training procedure. As an
alternative to this, we propose learning a set of complementary default values
from a large database of prior empirical results. Selecting an appropriate
configuration on a new dataset then requires only a simple, efficient and
embarrassingly parallel search over this set. We demonstrate the effectiveness
and efficiency of the approach we propose in comparison to random search and
Bayesian Optimization.
| stat |
Manifold lifting: scaling MCMC to the vanishing noise regime | Standard Markov chain Monte Carlo methods struggle to explore distributions
that are concentrated in the neighbourhood of low-dimensional structures. These
pathologies naturally occur in a number of situations. For example, they are
common to Bayesian inverse problem modelling and Bayesian neural networks, when
observational data are highly informative, or when a subset of the statistical
parameters of interest are non-identifiable. In this paper, we propose a
strategy that transforms the original sampling problem into the task of
exploring a distribution supported on a manifold embedded in a higher
dimensional space; in contrast to the original posterior this lifted
distribution remains diffuse in the vanishing noise limit. We employ a
constrained Hamiltonian Monte Carlo method which exploits the manifold geometry
of this lifted distribution, to perform efficient approximate inference. We
demonstrate in several numerical experiments that, contrarily to competing
approaches, the sampling efficiency of our proposed methodology does not
degenerate as the target distribution to be explored concentrates near low
dimensional structures.
| stat |
Stochastic Modeling of an Infectious Disease Part III-A: Analysis of
Time-Nonhomogeneous Models | We extend our BDI (birth-death-immigration) process based stochastic model of
an infectious disease to time-nonhomogeneous cases. First, we discuss the
deterministic model, and derive the expected value of the infection process.
Then as an application we consider that a government issues a decree to its
citizens to curtail their activities that may incur further infections and show
how the public's tardy response may further increase infections and prolong the
epidemic much longer than one might think. We seek to solve a partial
differential equation for the probability generating function. We find,
however, that an exact solution is obtainable only for the BD process, i.e., no
arrivals of the infected from outside. The coefficient of variation for the
nonhomogeneous BD process is found to be well over unity. This result implies
that the variations among different sample paths will be as large as in the
negative binomial distribution with r<1, which was found in Part I for the
homogeneous BDI model. In the final section, we illustrate, using our running
example, how much information we can derive from the time dependent PMF
(probability mass function) P_k(t)=Pr[I(t)=k]. We present graphical plots of
the PMF at various t's, and cross-sections of this function at various k's. A
mesh plot of the function over the (k, t) plane summarizes the above numerous
plots. The results of this paper reinforce our earlier claim (see Abstract of
Part II) that it would be a futile effort to attempt to identify all possible
reasons why environments of similar situations differ so much in their epidemic
patterns. Mere "luck" plays a more significant role than most of us may
believe. We should be prepared for a worse possible scenario, which only a
stochastic model can provide with probabilistic qualification. An empirical
validation of the above results will be given in Part III-B.
| stat |
Risk-averse estimation, an axiomatic approach to inference, and
Wallace-Freeman without MML | We define a new class of Bayesian point estimators, which we refer to as risk
averse. Using this definition, we formulate axioms that provide natural
requirements for inference, e.g. in a scientific setting, and show that for
well-behaved estimation problems the axioms uniquely characterise an estimator.
Namely, for estimation problems in which some parameter values have a positive
posterior probability (such as, e.g., problems with a discrete hypothesis
space), the axioms characterise Maximum A Posteriori (MAP) estimation, whereas
elsewhere (such as in continuous estimation) they characterise the
Wallace-Freeman estimator.
Our results provide a novel justification for the Wallace-Freeman estimator,
which previously was derived only as an approximation to the
information-theoretic Strict Minimum Message Length estimator. By contrast, our
derivation requires neither approximations nor coding.
| stat |
The Statistical Performance of Matching-Adjusted Indirect Comparisons | Indirect comparisons of treatment-specific outcomes across separate studies
often inform decision-making in the absence of head-to-head randomized
comparisons. Differences in baseline characteristics between study populations
may introduce confounding bias in such comparisons. Matching-adjusted indirect
comparison (MAIC) (Signorovitch et al., 2010) has been used to adjust for
differences in observed baseline covariates when the individual patient-level
data (IPD) are available for only one study and aggregate data (AGD) are
available for the other study. The approach weights outcomes from the IPD using
estimates of trial selection odds that balance baseline covariates between the
IPD and AGD. With the increasing use of MAIC, there is a need for formal
assessments of its statistical properties. In this paper we formulate
identification assumptions for causal estimands that justify MAIC estimators.
We then examine large sample properties and evaluate strategies for estimating
standard errors without the full IPD from both studies. The finite-sample bias
of MAIC and the performance of confidence intervals based on different standard
error estimators are evaluated through simulations. The method is illustrated
through an example comparing placebo arm and natural history outcomes in
Duchenne muscular dystrophy.
| stat |
An efficient Gehan-type estimation for the accelerated failure time
model with clustered and censored data | In medical studies, the collected covariates usually contain underlying
outliers. For clustered /longitudinal data with censored observations, the
traditional Gehan-type estimator is robust to outliers existing in response but
sensitive to outliers in the covariate domain, and it also ignores the
within-cluster correlations. To take account of within-cluster correlations,
varying cluster sizes, and outliers in covariates, we propose weighted
Gehan-type estimating functions for parameter estimation in the accelerated
failure time model for clustered data. We provide the asymptotic properties of
the resulting estimators and carry out simulation studies to evaluate the
performance of the proposed method under a variety of realistic settings. The
simulation results demonstrate that the proposed method is robust to the
outliers existing in the covariate domain and lead to much more efficient
estimators when a strong within-cluster correlation exists. Finally, the
proposed method is applied to a medical dataset and more reliable and
convincing results are hence obtained.
| stat |
A score function for Bayesian cluster analysis | We propose a score function for Bayesian clustering. The function is
parameter free and captures the interplay between the within cluster variance
and the between cluster entropy of a clustering. It can be used to choose the
number of clusters in well-established clustering methods such as hierarchical
clustering or $K$-means algorithm.
| stat |
A spatial Poisson hurdle model with application to wildfires | Modelling wildfire occurrences is important for disaster management including
prevention, detection and suppression of large catastrophic events. We present
a spatial Poisson hurdle model for exploring geographical variation of monthly
counts of wildfire occurrences and apply it to Indonesia and Australia. The
model includes two a priori independent spatially structured latent effects
that account for residual spatial variation in the probability of wildfire
occurrence, and the positive count rate given an occurrence. Inference is
provided by empirical Bayes using the Laplace approximation to the marginal
posterior which provides fast inference for latent Gaussian models with sparse
structures. In both cases, our model matched several empirically known facts
about wildfires. We conclude that elevation, percentage tree cover, relative
humidity, surface temperature, and the interaction between humidity and
temperature to be important predictors of monthly counts of wildfire
occurrences. Further, our findings show opposing effects for surface
temperature and its interaction with relative humidity.
| stat |
SubTSBR to tackle high noise and outliers for data-driven discovery of
differential equations | Data-driven discovery of differential equations has been an emerging research
topic. We propose a novel algorithm subsampling-based threshold sparse Bayesian
regression (SubTSBR) to tackle high noise and outliers. The subsampling
technique is used for improving the accuracy of the Bayesian learning
algorithm. It has two parameters: subsampling size and the number of
subsamples. When the subsampling size increases with fixed total sample size,
the accuracy of our algorithm goes up and then down. When the number of
subsamples increases, the accuracy of our algorithm keeps going up. We
demonstrate how to use our algorithm step by step and compare our algorithm
with threshold sparse Bayesian regression (TSBR) for the discovery of
differential equations. We show that our algorithm produces better results. We
also discuss the merits of discovering differential equations from data and
demonstrate how to discover models with random initial and boundary condition
as well as models with bifurcations. The numerical examples are: (1)
predator-prey model with noise, (2) shallow water equations with outliers, (3)
heat diffusion with random initial and boundary condition, and (4)
fish-harvesting problem with bifurcations.
| stat |
Neural networks in day-ahead electricity price forecasting: Single vs.
multiple outputs | Recent advancements in the fields of artificial intelligence and machine
learning methods resulted in a significant increase of their popularity in the
literature, including electricity price forecasting. Said methods cover a very
broad spectrum, from decision trees, through random forests to various
artificial neural network models and hybrid approaches. In electricity price
forecasting, neural networks are the most popular machine learning method as
they provide a non-linear counterpart for well-tested linear regression models.
Their application, however, is not straightforward, with multiple
implementation factors to consider. One of such factors is the network's
structure. This paper provides a comprehensive comparison of two most common
structures when using the deep neural networks -- one that focuses on each hour
of the day separately, and one that reflects the daily auction structure and
models vectors of the prices. The results show a significant accuracy advantage
of using the latter, confirmed on data from five distinct power exchanges.
| stat |
Smaller $p$-values via indirect information | This article develops $p$-values for evaluating means of normal populations
that make use of indirect or prior information. A $p$-value of this type is
based on a biased test statistic that is optimal on average with respect to a
probability distribution that encodes indirect information about the mean
parameter, resulting in a smaller $p$-value if the indirect information is
accurate. In a variety of multiparameter settings, we show how to adaptively
estimate the indirect information for each mean parameter while still
maintaining uniformity of the $p$-values under their null hypotheses. This is
done using a linking model through which indirect information about the mean of
one population may be obtained from the data of other populations. Importantly,
the linking model does not need to be correct to maintain the uniformity of the
$p$-values under their null hypotheses. This methodology is illustrated in
several data analysis scenarios, including small area inference, spatially
arranged populations, interactions in linear regression, and generalized linear
models.
| stat |
Penalized Projected Kernel Calibration for Computer Models | Projected kernel calibration is known to be theoretically superior, its loss
function is abbreviated as PK loss function. In this work, we prove the uniform
convergence of PK loss function and show that (1) when the sample size is
large, any local minimum point and local maximum point of the $L_2$ loss
between the true process and the computer models is a local minimum point of
the PK loss function; (2) all the local minimum values of the PK loss function
converge to the same value. These theoretical results imply that it is
extremely hard for the projected kernel calibration to identify the global
minimum point of the $L_2$ loss which is defined as the optimal value of the
calibration parameters. To solve this problem, a frequentist method, called the
penalized projected kernel calibration method is proposed. As a frequentist
method, the proposed method is proved to be semi-parametric efficient. On the
other hand, the proposed method has a natural bayesian version, which allows
users to calculate the credible region of the calibration parameters without
using a large sample approximation. Through extensive simulation studies and a
real-world case study, we show that the proposed calibration can accurately
estimate the calibration parameters, and compare favorably to alternative
calibration methods regardless of the sample size.
| stat |
Online stochastic gradient descent on non-convex losses from
high-dimensional inference | Stochastic gradient descent (SGD) is a popular algorithm for optimization
problems arising in high-dimensional inference tasks. Here one produces an
estimator of an unknown parameter from independent samples of data by
iteratively optimizing a loss function. This loss function is random and often
non-convex. We study the performance of the simplest version of SGD, namely
online SGD, from a random start in the setting where the parameter space is
high-dimensional.
We develop nearly sharp thresholds for the number of samples needed for
consistent estimation as one varies the dimension. Our thresholds depend only
on an intrinsic property of the population loss which we call the information
exponent. In particular, our results do not assume uniform control on the loss
itself, such as convexity or uniform derivative bounds. The thresholds we
obtain are polynomial in the dimension and the precise exponent depends
explicitly on the information exponent. As a consequence of our results, we
find that except for the simplest tasks, almost all of the data is used simply
in the initial search phase to obtain non-trivial correlation with the ground
truth. Upon attaining non-trivial correlation, the descent is rapid and
exhibits law of large numbers type behavior.
We illustrate our approach by applying it to a wide set of inference tasks
such as phase retrieval, and parameter estimation for generalized linear
models, online PCA, and spiked tensor models, as well as to supervised learning
for single-layer networks with general activation functions.
| stat |
Robust Graph Embedding with Noisy Link Weights | We propose $\beta$-graph embedding for robustly learning feature vectors from
data vectors and noisy link weights. A newly introduced empirical moment
$\beta$-score reduces the influence of contamination and robustly measures the
difference between the underlying correct expected weights of links and the
specified generative model. The proposed method is computationally tractable;
we employ a minibatch-based efficient stochastic algorithm and prove that this
algorithm locally minimizes the empirical moment $\beta$-score. We conduct
numerical experiments on synthetic and real-world datasets.
| stat |
A Binary Regression Adaptive Goodness-of-fit Test (BAGofT) | The Pearson's $\chi^2$ test and residual deviance test are two classical
goodness-of-fit tests for binary regression models such as logistic regression.
These two tests cannot be applied when we have one or more continuous
covariates in the data, a quite common situation in practice. In that case, the
most widely used approach is the Hosmer-Lemeshow test, which partitions the
covariate space into groups according to quantiles of the fitted probabilities
from all the observations. However, its grouping scheme is not flexible enough
to explore how to adversarially partition the data space in order to enhance
the power. In this work, we propose a new methodology, named binary regression
adaptive grouping goodness-of-fit test (BAGofT), to address the above concern.
It is a two-stage solution where the first stage adaptively selects candidate
partitions using "training" data, and the second stage performs $\chi^2$ tests
with necessary corrections based on "test" data. A proper data splitting
ensures that the test has desirable size and power properties. From our
experimental results, BAGofT performs much better than Hosmer-Lemeshow test in
many situations.
| stat |
Accelerography: Feasibility of Gesture Typing using Accelerometer | In this paper, we aim to look into the feasibility of constructing alphabets
using gestures. The main idea is to construct gestures, that are easy to
remember, not cumbersome to reproduce and easily identifiable. We construct
gestures for the entire English alphabet and provide an algorithm to identify
the gestures, even when they are constructed continuously. We tackle the
problem statistically, taking into account the problem of randomness in the
hand movement gestures of users, and achieve an average accuracy of 97.33% with
the entire English alphabet.
| stat |
On the Use of Random Forest for Two-Sample Testing | Following the line of classification-based two-sample testing, tests based on
the Random Forest classifier are proposed. The developed tests are easy to use,
require almost no tuning, and are applicable for any distribution on
$\mathbb{R}^d$. Furthermore, the built-in variable importance measure of the
Random Forest gives potential insights into which variables make out the
difference in distribution. An asymptotic power analysis for the proposed tests
is developed. Finally, two real-world applications illustrate the usefulness of
the introduced methodology. To simplify the use of the method, the R-package
"hypoRF" is provided.
| stat |
Sequential Tests of Multiple Hypotheses Controlling False Discovery and
Nondiscovery Rates | We propose a general and flexible procedure for testing multiple hypotheses
about sequential (or streaming) data that simultaneously controls both the
false discovery rate (FDR) and false nondiscovery rate (FNR) under minimal
assumptions about the data streams which may differ in distribution, dimension,
and be dependent. All that is needed is a test statistic for each data stream
that controls the conventional type I and II error probabilities, and no
information or assumptions are required about the joint distribution of the
statistics or data streams. The procedure can be used with sequential, group
sequential, truncated, or other sampling schemes. The procedure is a natural
extension of Benjamini and Hochberg's (1995) widely-used fixed sample size
procedure to the domain of sequential data, with the added benefit of
simultaneous FDR and FNR control that sequential sampling affords. We prove the
procedure's error control and give some tips for implementation in commonly
encountered testing situations.
| stat |
Clustering of variables for enhanced interpretability of predictive
models | A new strategy is proposed for building easy to interpret predictive models
in the context of a high-dimensional dataset, with a large number of highly
correlated explanatory variables. The strategy is based on a first step of
variables clustering using the CLustering of Variables around Latent Variables
(CLV) method. The exploration of the hierarchical clustering dendrogram is
undertaken in order to sequentially select the explanatory variables in a
group-wise fashion. For model setting implementation, the dendrogram is used as
the base-learner in an L2-boosting procedure. The proposed approach, named
lmCLV, is illustrated on the basis of a toy-simulated example when the clusters
and predictive equation are already known, and on a real case study dealing
with the authentication of orange juices based on 1H-NMR spectroscopic
analysis. In both illustrative examples, this procedure was shown to have
similar predictive efficiency to other methods, with additional
interpretability capacity. It is available in the R package ClustVarLV.
| stat |
Adversarial Alignment of Class Prediction Uncertainties for Domain
Adaptation | We consider unsupervised domain adaptation: given labelled examples from a
source domain and unlabelled examples from a related target domain, the goal is
to infer the labels of target examples. Under the assumption that features from
pre-trained deep neural networks are transferable across related domains,
domain adaptation reduces to aligning source and target domain at class
prediction uncertainty level. We tackle this problem by introducing a method
based on adversarial learning which forces the label uncertainty predictions on
the target domain to be indistinguishable from those on the source domain.
Pre-trained deep neural networks are used to generate deep features having high
transferability across related domains. We perform an extensive experimental
analysis of the proposed method over a wide set of publicly available
pre-trained deep neural networks. Results of our experiments on domain
adaptation tasks for image classification show that class prediction
uncertainty alignment with features extracted from pre-trained deep neural
networks provides an efficient, robust and effective method for domain
adaptation.
| stat |
Variance Estimation and Confidence Intervals from High-dimensional
Genome-wide Association Studies Through Misspecified Mixed Model Analysis | We study variance estimation and associated confidence intervals for
parameters characterizing genetic effects from genome-wide association studies
(GWAS) misspecified mixed model analysis. Previous studies have shown that, in
spite of the model misspecification, certain quantities of genetic interests
are estimable, and consistent estimators of these quantities can be obtained
using the restricted maximum likelihood (REML) method under a misspecified
linear mixed model. However, the asymptotic variance of such a REML estimator
is complicated and not ready to be implemented for practical use. In this
paper, we develop practical and computationally convenient methods for
estimating such asymptotic variances and constructing the associated confidence
intervals. Performance of the proposed methods is evaluated empirically based
on Monte-Carlo simulations and real-data application.
| stat |
Increasing the efficiency of Sequential Monte Carlo samplers through the
use of approximately optimal L-kernels | By facilitating the generation of samples from arbitrary probability
distributions, Markov Chain Monte Carlo (MCMC) is, arguably, \emph{the} tool
for the evaluation of Bayesian inference problems that yield non-standard
posterior distributions. In recent years, however, it has become apparent that
Sequential Monte Carlo (SMC) samplers have the potential to outperform MCMC in
a number of ways. SMC samplers are better suited to highly parallel computing
architectures and also feature various tuning parameters that are not available
to MCMC. One such parameter - the `L-kernel' - is a user-defined probability
distribution that can be used to influence the efficiency of the sampler. In
the current paper, the authors explain how to derive an expression for the
L-kernel that minimises the variance of the estimates realised by an SMC
sampler. Various approximation methods are then proposed to aid implementation
of the proposed L-kernel. The improved performance of the resulting algorithm
is demonstrated in multiple scenarios. For the examples shown in the current
paper, the use of an approximately optimum L-kernel has reduced the variance of
the SMC estimates by up to 99 % while also reducing the number of times that
resampling was required by between 65 % and 70 %. Python code and code tests
accompanying this manuscript are available through the Github repository
\url{https://github.com/plgreenLIRU/SMC_approx_optL}.
| stat |
Learning Optimal Tree Models Under Beam Search | Retrieving relevant targets from an extremely large target set under
computational limits is a common challenge for information retrieval and
recommendation systems. Tree models, which formulate targets as leaves of a
tree with trainable node-wise scorers, have attracted a lot of interests in
tackling this challenge due to their logarithmic computational complexity in
both training and testing. Tree-based deep models (TDMs) and probabilistic
label trees (PLTs) are two representative kinds of them. Though achieving many
practical successes, existing tree models suffer from the training-testing
discrepancy, where the retrieval performance deterioration caused by beam
search in testing is not considered in training. This leads to an intrinsic gap
between the most relevant targets and those retrieved by beam search with even
the optimally trained node-wise scorers. We take a first step towards
understanding and analyzing this problem theoretically, and develop the concept
of Bayes optimality under beam search and calibration under beam search as
general analyzing tools for this purpose. Moreover, to eliminate the
discrepancy, we propose a novel algorithm for learning optimal tree models
under beam search. Experiments on both synthetic and real data verify the
rationality of our theoretical analysis and demonstrate the superiority of our
algorithm compared to state-of-the-art methods.
| stat |
Learning Multimorbidity Patterns from Electronic Health Records Using
Non-negative Matrix Factorisation | Multimorbidity, or the presence of several medical conditions in the same
individual, has been increasing in the population, both in absolute and
relative terms. However, multimorbidity remains poorly understood, and the
evidence from existing research to describe its burden, determinants and
consequences has been limited. Previous studies attempting to understand
multimorbidity patterns are often cross-sectional and do not explicitly account
for multimorbidity patterns' evolution over time; some of them are based on
small datasets and/or use arbitrary and narrow age ranges; and those that
employed advanced models, usually lack appropriate benchmarking and
validations. In this study, we (1) introduce a novel approach for using
Non-negative Matrix Factorisation (NMF) for temporal phenotyping (i.e.,
simultaneously mining disease clusters and their trajectories); (2) provide
quantitative metrics for the evaluation of disease clusters from such studies;
and (3) demonstrate how the temporal characteristics of the disease clusters
that result from our model can help mine multimorbidity networks and generate
new hypotheses for the emergence of various multimorbidity patterns over time.
We trained and evaluated our models on one of the world's largest electronic
health records (EHR), with 7 million patients, from which over 2 million where
relevant to this study.
| stat |
Linear and Quadratic Discriminant Analysis: Tutorial | This tutorial explains Linear Discriminant Analysis (LDA) and Quadratic
Discriminant Analysis (QDA) as two fundamental classification methods in
statistical and probabilistic learning. We start with the optimization of
decision boundary on which the posteriors are equal. Then, LDA and QDA are
derived for binary and multiple classes. The estimation of parameters in LDA
and QDA are also covered. Then, we explain how LDA and QDA are related to
metric learning, kernel principal component analysis, Mahalanobis distance,
logistic regression, Bayes optimal classifier, Gaussian naive Bayes, and
likelihood ratio test. We also prove that LDA and Fisher discriminant analysis
are equivalent. We finally clarify some of the theoretical concepts with
simulations we provide.
| stat |
Scalable Gaussian Processes on Discrete Domains | Kernel methods on discrete domains have shown great promise for many
challenging data types, for instance, biological sequence data and molecular
structure data. Scalable kernel methods like Support Vector Machines may offer
good predictive performances but do not intrinsically provide uncertainty
estimates. In contrast, probabilistic kernel methods like Gaussian Processes
offer uncertainty estimates in addition to good predictive performance but fall
short in terms of scalability. We present the first sparse Gaussian Process
approximation framework on discrete input domains. Our framework achieves good
predictive performance as well as uncertainty estimates using discrete
optimization techniques. We present competitive results comparing our framework
to baseline methods such as Support Vector Machines and full Gaussian Processes
on synthetic data as well as on challenging real-world DNA sequence data.
| stat |
rcosmo: R Package for Analysis of Spherical, HEALPix and Cosmological
Data | The analysis of spatial observations on a sphere is important in areas such
as geosciences, physics and embryo research, just to name a few. The purpose of
the package rcosmo is to conduct efficient information processing,
visualisation, manipulation and spatial statistical analysis of Cosmic
Microwave Background (CMB) radiation and other spherical data. The package was
developed for spherical data stored in the Hierarchical Equal Area isoLatitude
Pixelation (Healpix) representation. rcosmo has more than 100 different
functions. Most of them initially were developed for CMB, but also can be used
for other spherical data as rcosmo contains tools for transforming spherical
data in cartesian and geographic coordinates into the HEALPix representation.
We give a general description of the package and illustrate some important
functionalities and benchmarks.
| stat |
Implementing version control with Git and GitHub as a learning objective
in statistics and data science courses | A version control system records changes to a file or set of files over time
so that changes can be tracked and specific versions of a file can be recalled
later. As such, it is an essential element of a reproducible workflow that
deserves due consideration among the learning objectives of statistics and data
science courses. This paper describes experiences and implementation decisions
of four contributing faculty who are teaching different courses at a variety of
institutions. Each of these faculty have set version control as a learning
objective and successfully integrated one such system (Git) into one or more
statistics courses. The various approaches described in the paper span
different implementation strategies to suit student background, course type,
software choices, and assessment practices. By presenting a wide range of
approaches to teaching Git, the paper aims to serve as a resource for
statistics and data science instructors teaching courses at any level within an
undergraduate or graduate curriculum.
| stat |
Surrogate-guided sampling designs for classification of rare outcomes
from electronic medical records data | Scalable and accurate identification of specific clinical outcomes has been
enabled by machine-learning applied to electronic medical record (EMR) systems.
The development of classification models requires the collection of a complete
labeled data set, where true clinical outcomes are obtained by human expert
manual review. For example, the development of natural language processing
algorithms requires the abstraction of clinical text data to obtain outcome
information necessary for training models. However, if the outcome is rare then
simple random sampling results in very few cases and insufficient information
to develop accurate classifiers. Since large scale detailed abstraction is
often expensive, time-consuming, and not feasible, more efficient strategies
are needed. Under such resource constrained settings, we propose a class of
enrichment sampling designs, where selection for abstraction is stratified by
auxiliary variables related to the true outcome of interest. Stratified
sampling on highly specific variables results in targeted samples that are more
enriched with cases, which we show translates to increased model discrimination
and better statistical learning performance. We provide mathematical details,
and simulation evidence that links sampling designs to their resulting
prediction model performance. We discuss the impact of our proposed sampling on
both model training and validation. Finally, we illustrate the proposed designs
for outcome label collection and subsequent machine-learning, using radiology
report text data from the Lumbar Imaging with Reporting of Epidemiology (LIRE)
study.
| stat |
Modeling dependent survival data through random effects with spatial
correlation at the subject level | Dynamical phenomena such as infectious diseases are often investigated by
following up subjects longitudinally, thus generating time to event data. The
spatial aspect of such data is also of primordial importance, as many
infectious diseases are transmitted from one subject to another. In this paper,
a spatially correlated frailty model is introduced that accommodates for the
correlation between subjects based on the distance between them. Estimates are
obtained through a stochastic approximation version of the Expectation
Maximization algorithm combined with a Monte-Carlo Markov Chain, for which
convergence is proven. The novelty of this model is that spatial correlation is
introduced for survival data at the subject level, each subject having its own
frailty. This univariate spatially correlated frailty model is used to analyze
spatially dependent malaria data, and its results are compared with other
standard models.
| stat |
Robust weights that optimally balance confounders for estimating
marginal hazard ratios | Covariate balance is crucial in obtaining unbiased estimates of treatment
effects in observational studies. Methods based on inverse probability weights
have been widely used to estimate treatment effects with observational data.
Machine learning techniques have been proposed to estimate propensity scores.
These techniques however target accuracy instead of covariate balance. Methods
that target covariate balance have been successfully proposed and largely
applied to estimate treatment effects on continuous outcomes. However, in many
medical and epidemiological applications, the interest lies in estimating
treatment effects on time-to-event outcomes. With this type of data, one of the
most common estimands of interest is the marginal hazard ratio of the Cox
proportional hazard model. In this paper, we start by presenting robust
orthogonality weights (ROW), a set of weights obtained by solving a quadratic
constrained optimization problem that maximizes precision while constraining
covariate balance defined as the sample correlation between confounders and
treatment. By doing so, ROW optimally deal with both binary and continuous
treatments. We then evaluate the performance of the proposed weights in
estimating marginal hazard ratios of binary and continuous treatments with
time-to-event outcomes in a simulation study. We finally apply ROW on the
evaluation of the effect of hormone therapy on time to coronary heart disease
and on the effect of red meat consumption on time to colon cancer among 24,069
postmenopausal women enrolled in the Women's Health Initiative observational
study.
| stat |
On the correspondence of deviances and maximum likelihood and interval
estimates from log-linear to logistic regression modelling | Consider a set of categorical variables $\mathcal{P}$ where at least one,
denoted by $Y$, is binary. The log-linear model that describes the counts in
the resulting contingency table implies a specific logistic regression model,
with the binary variable as the outcome. Extending results in Christensen
(1997), by also considering the case where factors present in the contingency
table disappear from the logistic regression model, we prove that the Maximum
Likelihood Estimate (MLE) for the parameters of the logistic regression equals
the MLE for the corresponding parameters of the log-linear model. We prove
that, asymptotically, standard errors for the two sets of parameters are also
equal. Subsequently, Wald confidence intervals are asymptotically equal. These
results demonstrate the extent to which inferences from the log-linear
framework can be translated to inferences within the logistic regression
framework, on the magnitude of main effects and interactions. Finally, we prove
that the deviance of the log-linear model is equal to the deviance of the
corresponding logistic regression, provided that the latter is fitted to a
dataset where no cell observations are merged when one or more factors in
$\mathcal{P} \setminus \{ Y \}$ become obsolete. We illustrate the derived
results with the analysis of a real dataset.
| stat |
A Statistician Teaches Deep Learning | Deep learning (DL) has gained much attention and become increasingly popular
in modern data science. Computer scientists led the way in developing deep
learning techniques, so the ideas and perspectives can seem alien to
statisticians. Nonetheless, it is important that statisticians become involved
-- many of our students need this expertise for their careers. In this paper,
developed as part of a program on DL held at the Statistical and Applied
Mathematical Sciences Institute, we address this culture gap and provide tips
on how to teach deep learning to statistics graduate students. After some
background, we list ways in which DL and statistical perspectives differ,
provide a recommended syllabus that evolved from teaching two iterations of a
DL graduate course, offer examples of suggested homework assignments, give an
annotated list of teaching resources, and discuss DL in the context of two
research areas.
| stat |
Robust adaptive variable selection in ultra-high dimensional linear
regression models | We consider the problem of simultaneous variable selection and estimation of
the corresponding regression coefficients in an ultra-high dimensional linear
regression models, an extremely important problem in the recent era. The
adaptive penalty functions are used in this regard to achieve the oracle
variable selection property along with easier computational burden. However,
the usual adaptive procedures (e.g., adaptive LASSO) based on the squared error
loss function is extremely non-robust in the presence of data contamination
which are quite common with large-scale data (e.g., noisy gene expression data,
spectra and spectral data). In this paper, we present a regularization
procedure for the ultra-high dimensional data using a robust loss function
based on the popular density power divergence (DPD) measure along with the
adaptive LASSO penalty. We theoretically study the robustness and the
large-sample properties of the proposed adaptive robust estimators for a
general class of error distributions; in particular, we show that the proposed
adaptive DPD-LASSO estimator is highly robust, satisfies the oracle variable
selection property, and the corresponding estimators of the regression
coefficients are consistent and asymptotically normal under easily verifiable
set of assumptions. Numerical illustrations are provided for the mostly used
normal error density. Finally, the proposal is applied to analyze an
interesting spectral dataset, in the field of chemometrics, regarding the
electron-probe X-ray microanalysis (EPXMA) of archaeological glass vessels from
the 16th and 17th centuries.
| stat |
Point Process Modeling of Drug Overdoses with Heterogeneous and Missing
Data | Opioid overdose rates have increased in the United States over the past
decade and reflect a major public health crisis. Modeling and prediction of
drug and opioid hotspots, where a high percentage of events fall in a small
percentage of space-time, could help better focus limited social and health
services. In this work we present a spatial-temporal point process model for
drug overdose clustering. The data input into the model comes from two
heterogeneous sources: 1) high volume emergency medical calls for service (EMS)
records containing location and time, but no information on the type of
non-fatal overdose and 2) fatal overdose toxicology reports from the coroner
containing location and high-dimensional information from the toxicology screen
on the drugs present at the time of death. We first use non-negative matrix
factorization to cluster toxicology reports into drug overdose categories and
we then develop an EM algorithm for integrating the two heterogeneous data
sets, where the mark corresponding to overdose category is inferred for the EMS
data and the high volume EMS data is used to more accurately predict drug
overdose death hotspots. We apply the algorithm to drug overdose data from
Indianapolis, showing that the point process defined on the integrated data
outperforms point processes that use only homogeneous EMS (AUC improvement .72
to .8) or coroner data (AUC improvement .81 to .85).We also investigate the
extent to which overdoses are contagious, as a function of the type of
overdose, while controlling for exogenous fluctuations in the background rate
that might also contribute to clustering. We find that drug and opioid overdose
deaths exhibit significant excitation, with branching ratio ranging from .72 to
.98.
| stat |
CDSM -- Casual Inference using Deep Bayesian Dynamic Survival Models | Causal inference in longitudinal observational health data often requires the
accurate estimation of treatment effects on time-to-event outcomes in the
presence of time-varying covariates. To tackle this sequential treatment effect
estimation problem, we have developed a causal dynamic survival model (CDSM)
that uses the potential outcomes framework with the Bayesian recurrent
sub-networks to estimate the difference in survival curves. Using simulated
survival datasets, CDSM has shown good causal effect estimation performance
across scenarios of sample dimension, event rate, confounding and overlapping.
However, we found increasing the sample size is not effective if the original
data is highly confounded or with low level of overlapping. In two large
clinical cohort studies, our model identified the expected conditional average
treatment effect and detected individual effect heterogeneity over time and
patient subgroups. The model provides individualized absolute treatment effect
estimations that could be used in recommendation systems.
| stat |
Detecting anomalies in fibre systems using 3-dimensional image data | We consider the problem of detecting anomalies in the directional
distribution of fibre materials observed in 3D images. We divide the image into
a set of scanning windows and classify them into two clusters: homogeneous
material and anomaly. Based on a sample of estimated local fibre directions,
for each scanning window we compute several classification attributes, namely
the coordinate wise means of local fibre directions, the entropy of the
directional distribution, and a combination of them. We also propose a new
spatial modification of the Stochastic Approximation Expectation-Maximization
(SAEM) algorithm. Besides the clustering we also consider testing the
significance of anomalies. To this end, we apply a change point technique for
random fields and derive the exact inequalities for tail probabilities of a
test statistics. The proposed methodology is first validated on simulated
images. Finally, it is applied to a 3D image of a fibre reinforced polymer.
| stat |
Discriminative Topic Modeling with Logistic LDA | Despite many years of research into latent Dirichlet allocation (LDA),
applying LDA to collections of non-categorical items is still challenging. Yet
many problems with much richer data share a similar structure and could benefit
from the vast literature on LDA. We propose logistic LDA, a novel
discriminative variant of latent Dirichlet allocation which is easy to apply to
arbitrary inputs. In particular, our model can easily be applied to groups of
images, arbitrary text embeddings, and integrates well with deep neural
networks. Although it is a discriminative model, we show that logistic LDA can
learn from unlabeled data in an unsupervised manner by exploiting the group
structure present in the data. In contrast to other recent topic models
designed to handle arbitrary inputs, our model does not sacrifice the
interpretability and principled motivation of LDA.
| stat |
Bayesian detection of piecewise linear trends in replicated time-series
with application to growth data modelling | We consider the situation where a temporal process is composed of contiguous
segments with differing slopes and replicated noise-corrupted time series
measurements are observed. The unknown mean of the data generating process is
modelled as a piecewise linear function of time with an unknown number of
change-points. We develop a Bayesian approach to infer the joint posterior
distribution of the number and position of change-points as well as the unknown
mean parameters. A-priori, the proposed model uses an overfitting number of
mean parameters but, conditionally on a set of change-points, only a subset of
them influences the likelihood. An exponentially decreasing prior distribution
on the number of change-points gives rise to a posterior distribution
concentrating on sparse representations of the underlying sequence. A
Metropolis-Hastings Markov chain Monte Carlo (MCMC) sampler is constructed for
approximating the posterior distribution. Our method is benchmarked using
simulated data and is applied to uncover differences in the dynamics of fungal
growth from imaging time course data collected from different strains. The
source code is available on CRAN.
| stat |
Stanza: A Nonlinear State Space Model for Probabilistic Inference in
Non-Stationary Time Series | Time series with long-term structure arise in a variety of contexts and
capturing this temporal structure is a critical challenge in time series
analysis for both inference and forecasting settings. Traditionally, state
space models have been successful in providing uncertainty estimates of
trajectories in the latent space. More recently, deep learning, attention-based
approaches have achieved state of the art performance for sequence modeling,
though often require large amounts of data and parameters to do so. We propose
Stanza, a nonlinear, non-stationary state space model as an intermediate
approach to fill the gap between traditional models and modern deep learning
approaches for complex time series. Stanza strikes a balance between
competitive forecasting accuracy and probabilistic, interpretable inference for
highly structured time series. In particular, Stanza achieves forecasting
accuracy competitive with deep LSTMs on real-world datasets, especially for
multi-step ahead forecasting.
| stat |
Conditional Sampling With Monotone GANs | We present a new approach for sampling conditional probability measures,
enabling consistent uncertainty quantification in supervised learning tasks. We
construct a mapping that transforms a reference measure to the measure of the
output conditioned on new inputs. The mapping is trained via a modification of
generative adversarial networks (GANs), called monotone GANs, that imposes
monotonicity and a block triangular structure. We present theoretical
guarantees for the consistency of our proposed method, as well as numerical
experiments demonstrating the ability of our method to accurately sample
conditional measures in applications ranging from inverse problems to image
in-painting.
| stat |
Task Agnostic Continual Learning Using Online Variational Bayes | Catastrophic forgetting is the notorious vulnerability of neural networks to
the change of the data distribution while learning. This phenomenon has long
been considered a major obstacle for allowing the use of learning agents in
realistic continual learning settings. A large body of continual learning
research assumes that task boundaries are known during training. However,
research for scenarios in which task boundaries are unknown during training has
been lacking. In this paper we present, for the first time, a method for
preventing catastrophic forgetting (BGD) for scenarios with task boundaries
that are unknown during training --- task-agnostic continual learning. Code of
our algorithm is available at https://github.com/igolan/bgd.
| stat |
The relative efficiency of time-to-progression and continuous measures
of cognition in pre-symptomatic Alzheimer's | Pre-symptomatic (or Preclinical) Alzheimer's Disease is defined by biomarker
evidence of fibrillar amyloid beta pathology in the absence of clinical
symptoms. Clinical trials in this early phase of disease are challenging due to
the slow rate of disease progression as measured by periodic cognitive
performance tests or by transition to a diagnosis of Mild Cognitive Impairment.
In a multisite study, experts provide diagnoses by central chart review without
the benefit of in-person assessment. We use a simulation study to demonstrate
that models of repeated cognitive assessments detect treatment effects more
efficiently compared to models of time-to-progression to an endpoint such as
change in diagnosis. Multivariate continuous data are simulated from a Bayesian
joint mixed effects model fit to data from the Alzheimer's Disease Neuroimaging
Initiative. Simulated progression events are algorithmically derived from the
continuous assessments using a random forest model fit to the same data. We
find that power is approximately doubled with models of repeated continuous
outcomes compared to the time-to-progression analysis. The simulations also
demonstrate that a plausible informative missing data pattern can induce a bias
which inflates treatment effects, yet 5% Type I error is maintained.
| stat |
A New Graphical Device and Related Tests for the Shape of Non-parametric
Regression Function | We consider a non-parametric regression model $y = m(x) + \epsilon$ and
propose a novel graphical device to check whether the $r$-th ($r \geqslant 1$)
derivative of the regression function $m(x)$ is positive or otherwise. Since
the shape of the regression function can be completely characterized by its
derivatives, the graphical device can correctly identify the shape of the
regression function. The proposed device includes the check for monotonicity
and convexity of the function as special cases. We also present an example to
elucidate the practical utility of the graphical device. In addition, we employ
the graphical device to formulate a class of test statistics and derive its
asymptotic distribution. The tests are exhibited in various simulated and real
data examples.
| stat |
Estimation of Climbing Route Difficulty using Whole-History Rating | Existing grading systems for rock climbing routes assign a difficulty grade
to a route based on the opinions of a few people. An objective approach to
estimating route difficulty on an interval scale was obtained by adapting the
Whole-History Rating (WHR) system to rock climbing. WHR's model was fitted to a
database of 236,095 ascents recorded by users on an established climbing
website. 73% of the ascents used in the dataset were classified as successful.
Predictions were on average 85% accurate with 10-fold cross-validation. The
results suggest that an empirical rating system is accurate at assessing route
difficulty and is viable for revising conventional route grades.
| stat |
Meta Analysis of Bayes Factors | Bayes Factors, the Bayesian tool for hypothesis testing, are receiving
increasing attention in the literature. Compared to their frequentist rivals
($p$-values or test statistics), Bayes Factors have the conceptual advantage of
providing evidence both for and against a null hypothesis and they can be
calibrated so that they do not depend so heavily on the sample size. However,
research on the synthesis of Bayes Factors arising from individual studies has
received very limited attention. In this work we review and propose methods for
combining Bayes Factors from multiple studies, depending on the level of
information available. In the process, we provide insights with respect to the
interplay between frequentist and Bayesian evidence. We also clarify why some
intuitive suggestions in the literature can be misleading. We assess the
performance of the methods discussed via a simulation study and apply the
methods in an example from the field of psychology.
| stat |
Application of a new information priority accumulated grey model with
time power to predict short-term wind turbine capacity | Wind energy makes a significant contribution to global power generation.
Predicting wind turbine capacity is becoming increasingly crucial for cleaner
production. For this purpose, a new information priority accumulated grey model
with time power is proposed to predict short-term wind turbine capacity.
Firstly, the computational formulas for the time response sequence and the
prediction values are deduced by grey modeling technique and the definite
integral trapezoidal approximation formula. Secondly, an intelligent algorithm
based on particle swarm optimization is applied to determine the optimal
nonlinear parameters of the novel model. Thirdly, three real numerical examples
are given to examine the accuracy of the new model by comparing with six
existing prediction models. Finally, based on the wind turbine capacity from
2007 to 2017, the proposed model is established to predict the total wind
turbine capacity in Europe, North America, Asia, and the world. The numerical
results reveal that the novel model is superior to other forecasting models. It
has a great advantage for small samples with new characteristic behaviors.
Besides, reasonable suggestions are put forward from the standpoint of the
practitioners and governments, which has high potential to advance the
sustainable improvement of clean energy production in the future.
| stat |
Divide-and-Conquer MCMC for Multivariate Binary Data | We analyze a large database of de-identified Medicare Advantage claims from a
single large US health insurance provider, where the number of individuals
available for analysis are an order of magnitude larger than the number of
potential covariates. This type of data, dubbed `tall data', often does not fit
in memory, and estimating parameters using traditional Markov Chain Monte Carlo
(MCMC) methods is a computationally infeasible task. We show how
divide-and-conquer MCMC, which splits the data into disjoint subsamples and
runs a MCMC algorithm on each sample in parallel before combining results, can
be used with a multivariate probit factor model. We then show how this approach
can be applied to large medical datasets to provide insights into questions of
interest to the medical community. We also conduct a simulation study,
comparing two posterior combination algorithms with a mean-field stochastic
variational approach, showing that divide-and-conquer MCMC should be preferred
over variational inference when estimating the latent correlation structure
between binary responses is of primary interest.
| stat |
Phylogenetically informed Bayesian truncated copula graphical models for
microbial association networks | Microorganisms play a critical role in host health. The advancement of
high-throughput sequencing technology provides opportunities for a deeper
understanding of microbial interactions. However, due to the limitations of 16S
ribosomal RNA sequencing, microbiome data are zero-inflated, and a quantitative
comparison of microbial abundances cannot be made across subjects. By
leveraging a recent microbiome profiling technique that quantifies 16S
ribosomal RNA microbial counts, we propose a novel Bayesian graphical model
that incorporates microorganisms' evolutionary history through a phylogenetic
tree prior and explicitly accounts for zero-inflation using the truncated
Gaussian copula. Our simulation study reveals that the evolutionary information
substantially improves the network estimation accuracy. We apply the proposed
model to the quantitative gut microbiome data of 106 healthy subjects, and
identify three distinct microbial communities that are not determined by
existing microbial network estimation models. We further find that these
communities are discriminated based on microorganisms' ability to utilize
oxygen as an energy source.
| stat |
Investigating Sprawl using AIC and Recursive Partitioning Trees: A
Machine Learning Approach to Assessing the Association between Poverty and
Commute Time | Sprawl, according to Glaeser and Kahn, is the 21st century phenomenon that
some people are not dependent on city-living due to automobiles and therefore
can live outside public transportation spheres and cities. This is usually seen
as pleasant and accompanied by improved qualities of life, but as they
addressed, the problem remains that sprawl causes loss of jobs for those who
cannot afford luxurious alternatives but only inferior substitutes (Glaeser and
Kahn 2004). Therefore, through our question, we hope to suggest that sprawl has
occurred in the U.S. and poverty is one of the consequences.
| stat |
A Data-Driven Approach for Discovering Stochastic Dynamical Systems with
Non-Gaussian Levy Noise | With the rapid increase of valuable observational, experimental and
simulating data for complex systems, great efforts are being devoted to
discovering governing laws underlying the evolution of these systems. However,
the existing techniques are limited to extract governing laws from data as
either deterministic differential equations or stochastic differential
equations with Gaussian noise. In the present work, we develop a new
data-driven approach to extract stochastic dynamical systems with non-Gaussian
symmetric L\'evy noise, as well as Gaussian noise. First, we establish a
feasible theoretical framework, by expressing the drift coefficient, diffusion
coefficient and jump measure (i.e., anomalous diffusion) for the underlying
stochastic dynamical system in terms of sample paths data. We then design a
numerical algorithm to compute the drift, diffusion coefficient and jump
measure, and thus extract a governing stochastic differential equation with
Gaussian and non-Gaussian noise. Finally, we demonstrate the efficacy and
accuracy of our approach by applying to several prototypical one-, two- and
three-dimensional systems. This new approach will become a tool in discovering
governing dynamical laws from noisy data sets, from observing or simulating
complex phenomena, such as rare events triggered by random fluctuations with
heavy as well as light tail statistical features.
| stat |
GSSMD: New metric for robust and interpretable assay quality assessment
and hit selection | In the high-throughput screening (HTS) campaigns, the Z'-factor and strictly
standardized mean difference (SSMD) are commonly used to assess the quality of
assays and to select hits. However, these measures are vulnerable to outliers
and their performances are highly sensitive to background distributions. Here,
we propose an alternative measure for assay quality assessment and hit
selection. The proposed method is a non-parametric generalized variant of SSMD
(GSSMD). In this paper, we have shown that the proposed method provides more
robust and intuitive way of assay quality assessment and hit selection.
| stat |
A Generalization of Spatial Monte Carlo Integration | Spatial Monte Carlo integration (SMCI) is an extension of standard Monte
Carlo integration and can approximate expectations on Markov random fields with
high accuracy. SMCI was applied to pairwise Boltzmann machine (PBM) learning,
with superior results to those from some existing methods. The approximation
level of SMCI can be changed, and it was proved that a higher-order
approximation of SMCI is statistically more accurate than a lower-order
approximation. However, SMCI as proposed in the previous studies suffers from a
limitation that prevents the application of a higher-order method to dense
systems.
This study makes two different contributions as follows. A generalization of
SMCI (called generalized SMCI (GSMCI)) is proposed, which allows relaxation of
the above-mentioned limitation; moreover, a statistical accuracy bound of GSMCI
is proved. This is the first contribution of this study. A new PBM learning
method based on SMCI is proposed, which is obtained by combining SMCI and the
persistent contrastive divergence. The proposed learning method greatly
improves the accuracy of learning. This is the second contribution of this
study.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.