Computational Statistics & Data Analysis最新文献

The number of modes in a probability density function is representative of the complexity of a model and can also be viewed as the number of subpopulations. Despite its relevance, there has been limited research in this area. A novel approach to estimating the number of modes in the univariate setting is presented, focusing on prediction accuracy and inspired by some overlooked aspects of the problem: the need for structure in the solutions, the subjective and uncertain nature of modes, and the convenience of a holistic view that blends local and global density properties. The technique combines flexible kernel estimators and parsimonious compositional splines in the Bayesian inference paradigm, providing soft solutions and incorporating expert judgment. The procedure includes feature exploration, model selection, and mode testing, illustrated in a sports analytics case study showcasing multiple companion visualisation tools. A thorough simulation study also demonstrates that traditional modality-driven approaches paradoxically struggle to provide accurate results. In this context, the new method emerges as a top-tier alternative, offering innovative solutions for analysts.

Bayesian inference with deep generative prior has received considerable interest for solving imaging inverse problems in many scientific and engineering fields. The selection of the prior distribution is learned from, and therefore an important representation learning of, available prior measurements. The SA-Roundtrip, a novel deep generative prior, is introduced to enable controlled sampling generation and identify the data's intrinsic dimension. This prior incorporates a self-attention structure within a bidirectional generative adversarial network. Subsequently, Bayesian inference is applied to the posterior distribution in the low-dimensional latent space using the Hamiltonian Monte Carlo with preconditioned Crank-Nicolson (HMC-pCN) algorithm, which is proven to be ergodic under specific conditions. Experiments conducted on computed tomography (CT) reconstruction with the MNIST and TomoPhantom datasets reveal that the proposed method outperforms state-of-the-art comparisons, consistently yielding a robust and superior point estimator along with precise uncertainty quantification.

Graphical models allow modeling of complex dependencies among components of a random vector. In many applications of graphical models, however, for example microbiome data, the data have an excess number of zero values. New pairwise graphical models with distributions in an exponential family are presented, that accommodate excess numbers of zeros in the random vector components. First these multivariate distributions are characterized in terms of univariate conditional distributions. Then predictors that arise from such a pairwise graphical model with excess zeros are modeled as functions of an outcome, and the corresponding first order sufficient dimension reduction (SDR) is derived. That is, linear combinations of the predictors that contain all the information for the regression of the outcome as a function of the predictors are obtained. To incorporate variable selection, the SDR is estimated using a pseudo-likelihood with a hierarchical penalty that prioritizes sparse interactions only for variables associated with the outcome. These methods yield consistent estimators of the reduction and can be applied to continuous or categorical outcomes. The new methods are then illustrated by studying normal, Poisson and truncated Poisson graphical models with excess zeros in simulations and by analyzing microbiome data from the American Gut Project. The models provided robust variable selection and the predictive performance of the Poisson zero-inflated pairwise graphical model was equal or better than that of other available methods for the analysis of microbiome data.

The logistic normal multinomial distribution is gaining interest in modelling microbiome data. It utilizes a hierarchical structure such that the observed counts conditional on the compositions are assumed to be multinomial random variables and the log-ratio transformed compositions are assumed to be from a Gaussian distribution. While multinomial distribution accounts for the compositional nature of the data, and a Gaussian prior offers flexibility in the structure of covariance matrices, the log-ratio transformed compositions of the microbiome data can be highly skewed, especially at a lower taxonomic level. Thus, a Gaussian distribution may not be an ideal prior for the log-ratio transformed compositions. A novel mixture of logistic skew-normal multinomial (LSNM) distribution is proposed in which a multivariate skew-normal distribution is utilized as a prior for the log-ratio transformed compositions. A variational Gaussian approximation in conjunction with the EM algorithm is utilized for parameter estimation.

Accelerated failure time (AFT) models with random effects, a useful alternative to frailty models, have been widely used for analyzing clustered (or correlated) time-to-event data. In the AFT model, the distribution of the unobserved random effect is conventionally assumed to be parametric, often modeled as a normal distribution. Although it has been known that a misspecfied random-effect distribution has little effect on regression parameter estimates, in some cases, the impact caused by such misspecification is not negligible. Particularly when our focus extends to quantities associated with random effects, the problem could become worse. In this paper, we propose a semi-parametric maximum likelihood approach in which the random-effect distribution under the AFT models is left unspecified. We provide a feasible algorithm to estimate the random-effect distribution as well as model parameters. Through comprehensive simulation studies, our results demonstrate the effectiveness of this proposed method across a range of random-effect distribution types (discrete or continuous) and under conditions of heavy censoring. The efficacy of the approach is further illustrated through simulation studies and real-world data examples.

Interval-censored failure time data frequently occur in medical follow-up studies among others and include right-censored data as a special case. Their analysis is much difficult than the analysis of the right-censored data due to their much more complicated structures and no partial likelihood. This article presents a variational Bayesian (VB) approach for analyzing such data under a proportional hazards model. The VB approach obtains a direct approximation of the posterior density. Compared to the Markov chain Monte Carlo (MCMC)-based sampling approaches, the VB approach achieves enhanced computational efficiency without sacrificing estimation accuracy. An extensive simulation study is conducted to compare the performance of the proposed methods with two main Bayesian methods currently available in the literature and the classic proportional hazards model and indicates that they work well in practical situations.

In multiple instance learning (MIL), a bag represents a sample that has a set of instances, each of which is described by a vector of explanatory variables, but the entire bag only has one label/response. Though many methods for MIL have been developed to date, few have paid attention to interpretability of models and results. The proposed Bayesian regression model stands on two levels of hierarchy, which transparently show how explanatory variables explain and instances contribute to bag responses. Moreover, two selection problems are simultaneously addressed; the instance selection to find out the instances in each bag responsible for the bag response, and the variable selection to search for the important covariates. To explore a joint discrete space of indicator variables created for selection of both explanatory variables and instances, the shotgun stochastic search algorithm is modified to fit in the MIL context. Also, the proposed model offers a natural and rigorous way to quantify uncertainty in coefficient estimation and outcome prediction, which many modern MIL applications call for. The simulation study shows the proposed regression model can select variables and instances with high performance (AUC greater than 0.86), thus predicting responses well. The proposed method is applied to the musk data for prediction of binding strengths (labels) between molecules (bags) with different conformations (instances) and target receptors. It outperforms all existing methods, and can identify variables relevant in modeling responses.

Multi-task learning (MTL) is a methodology that aims to improve the general performance of estimation and prediction by sharing common information among related tasks. In the MTL, there are several assumptions for the relationships and methods to incorporate them. One of the natural assumptions in the practical situation is that tasks are classified into some clusters with their characteristics. For this assumption, the group fused regularization approach performs clustering of the tasks by shrinking the difference among tasks. This enables the transfer of common information within the same cluster. However, this approach also transfers the information between different clusters, which worsens the estimation and prediction. To overcome this problem, an MTL method is proposed with a centroid parameter representing a cluster center of the task. Because this model separates parameters into the parameters for regression and the parameters for clustering, estimation and prediction accuracy for regression coefficient vectors are improved. The effectiveness of the proposed method is shown through Monte Carlo simulations and applications to real data.

The maximum likelihood problem for Hidden Markov Models is usually numerically solved by the Baum-Welch algorithm, which uses the Expectation-Maximization algorithm to find the estimates of the parameters. This algorithm has a recursion depth equal to the data sample size and cannot be computed in parallel, which limits the use of modern GPUs to speed up computation time. A new algorithm is proposed that provides the same estimates as the Baum-Welch algorithm, requiring about the same number of iterations, but is designed in such a way that it can be parallelized. As a consequence, it leads to a significant reduction in the computation time. This reduction is illustrated by means of numerical examples, where we consider simulated data as well as real datasets.

Semiparametric Cox proportional hazards models enjoy great popularity in empirical survival analysis. A semiparametric model for cause-specific hazards under a proportionality restriction across risks is considered, which has desired practical properties such as estimation by partial likelihood and an analytical solution for the copula-graphic estimator. The cause-specific and marginal hazards are shown to share functional form restrictions in this case. The model for the cause-specific hazard can be used for inference about parametric restrictions on the marginal hazard without the risk of misspecifying the latter and without knowing the risk dependence. After the class of parametric marginal hazards has been determined, it can be estimated in conjunction with the degree of risk dependence. Finite sample properties are investigated with simulations. An application to employment duration demonstrates the practicality of the approach.