arXiv subjects

Jeffrey S. Simonoff

Publications and source records attributed to Jeffrey S. Simonoff.

12 recordsLinked to original sources

arXiv2023

A Regression Tree Method for Longitudinal and Clustered Data with Multivariate Responses

RE-EM tree is a tree-based method that combines the regression tree and the linear mixed effects model for modeling univariate response longitudinal or clustered data. In this paper, we generalize the RE-EM tree method to multivariate response data, by adopting the Multivariate Regression Tree method proposed by De'Ath [2002]. The Multivariate RE-EM tree method estimates a population-level single tree structure that is driven by the multiple responses simultaneously and object-level random effects for each response variable, where correlation between the response variables and between the associated random effects are each allowed. Through simulation studies, we verify the advantage of the Multivariate RE-EM tree over the use of multiple univariate RE-EM trees and the Multivariate Regression Tree. We apply the Multivariate RE-EM tree to analyze a real data set that contains multidimensional nonfinancial characteristics of poverty of different countries as responses, and various potential causes of poverty as predictors.

stat.ME↗

arXiv2022

Ensemble methods for survival function estimation with time-varying covariates

Survival data with time-varying covariates are common in practice. If relevant, they can improve on the estimation of survival function. However, the traditional survival forests - conditional inference forest, relative risk forest and random survival forest - have accommodated only time-invariant covariates. We generalize the conditional inference and relative risk forests to allow time-varying covariates. We also propose a general framework for estimation of a survival function in the presence of time-varying covariates. We compare their performance with that of the Cox model and transformation forest, adapted here to accommodate time-varying covariates, through a comprehensive simulation study in which the Kaplan-Meier estimate serves as a benchmark, and performance is compared using the integrated L2 difference between the true and estimated survival functions. In general, the performance of the two proposed forests substantially improves over the Kaplan-Meier estimate. Taking into account all other factors, under the proportional hazard (PH) setting, the best method is always one of the two proposed forests, while under the non-PH setting, it is the adapted transformation forest. K-fold cross-validation is used as an effective tool to choose between the methods in practice.

stat.AP↗

arXiv2021

On the Use of Information Criteria for Subset Selection in Least Squares Regression

Least squares (LS)-based subset selection methods are popular in linear regression modeling. Best subset selection (BS) is known to be NP-hard and has a computational cost that grows exponentially with the number of predictors. Recently, Bertsimas (2016) formulated BS as a mixed integer optimization (MIO) problem and largely reduced the computation overhead by using a well-developed optimization solver, but the current methodology is not scalable to very large datasets. In this paper, we propose a novel LS-based method, the best orthogonalized subset selection (BOSS) method, which performs BS upon an orthogonalized basis of ordered predictors and scales easily to large problem sizes. Another challenge in applying LS-based methods in practice is the selection rule to choose the optimal subset size k. Cross-validation (CV) requires fitting a procedure multiple times, and results in a selected k that is random across repeated application to the same dataset. Compared to CV, information criteria only require fitting a procedure once, but they require knowledge of the effective degrees of freedom for the fitting procedure, which is generally not available analytically for complex methods. Since BOSS uses orthogonalized predictors, we first explore a connection for orthogonal non-random predictors between BS and its Lagrangian formulation (i.e., minimization of the residual sum of squares plus the product of a regularization parameter and k), and based on this connection propose a heuristic degrees of freedom (hdf) for BOSS that can be estimated via an analytically-based expression. We show in both simulations and real data analysis that BOSS using a proposed Kullback-Leibler based information criterion AICc-hdf has the strongest performance of all of the LS-based methods considered and is competitive with regularization methods, with the computational effort of a single ordinary LS fit.

stat.ME↗

arXiv2021

Dynamic estimation with random forests for discrete-time survival data

Time-varying covariates are often available in survival studies and estimation of the hazard function needs to be updated as new information becomes available. In this paper, we investigate several different easy-to-implement ways that random forests can be used for dynamic estimation of the survival or hazard function from discrete-time survival data. The results from a simulation study indicate that all methods can perform well, and that none dominates the others. In general, situations that are more difficult from an estimation point of view (such as weaker signals and less data) favour a global fit, pooling over all time points, while situations that are easier from an estimation point of view (such as stronger signals and more data) favor local fits.

stat.ME↗

arXiv2020

Joint Latent Class Trees: A Tree-Based Approach to Modeling Time-to-event and Longitudinal Data

In this paper, we propose a semiparametric, tree based joint latent class modeling approach (JLCT) to model the joint behavior of longitudinal and time-to-event data. Existing joint latent class modeling approaches are parametric and can suffer from high computational cost. The most common parametric approach, the joint latent class model (JLCM), further restricts analysis to using time-invariant covariates in modeling survival risks and latent class memberships. Instead, the proposed JLCT is fast to fit, and can use time-varying covariates in all of its modeling components. We demonstrate the prognostic value of using time-varying covariates, and therefore the advantage of JLCT over JLCM on simulated data. We further apply JLCT to the PAQUID data set and confirm its superior prediction performance and orders-of-magnitude speedup over JLCM.

stat.ME↗

arXiv2020

Selection of Regression Models under Linear Restrictions for Fixed and Random Designs

Many important modeling tasks in linear regression, including variable selection (in which slopes of some predictors are set equal to zero) and simplified models based on sums or differences of predictors (in which slopes of those predictors are set equal to each other, or the negative of each other, respectively), can be viewed as being based on imposing linear restrictions on regression parameters. In this paper, we discuss how such models can be compared using information criteria designed to estimate predictive measures like squared error and Kullback-Leibler (KL) discrepancy, in the presence of either deterministic predictors (fixed-X) or random predictors (random-X). We extend the justifications for existing fixed-X criteria Cp, FPE and AICc, and random-X criteria Sp and RCp, to general linear restrictions. We further propose and justify a KL-based criterion, RAICc, under random-X for variable selection and general linear restrictions. We show in simulations that the use of the KL-based criteria AICc and RAICc results in better predictive performance and sparser solutions than the use of squared error-based criteria, including cross-validation.

stat.ME↗

arXiv2020

Hot Racquet or Not? An Exploration of Momentum in Grand Slam Tennis Matches

The presence of momentum in sports, where the outcome of a previous event affects a following event, is a belief often held by fans, and has been the subject of statistical study for many sports. This paper investigates the presence and manifestation of momentum in Grand Slam tennis matches from 2014 through 2019 for men and women, to see if there is evidence of any carryover effect from the outcome of previous point(s)/game(s)/set(s) to the current one that cannot be accounted for by player quality, fitness or fatigue. Generalized linear mixed effect models (GLMMs) are used to explore the effects of the outcomes of previous sets, games, or points on the odds of winning a set, game, or point, while incorporating control variables that account for differences in player quality and the current status of the match. We find strong evidence of carryover effects at the set, game, and point level. Holding one's serve in prior service games is strongly related to winning a current game, but losing a past game is associated with higher estimated odds of winning the next game. Winning the previous two or three points in a row is associated with highest estimated odds of winning the next point.

stat.AP↗

arXiv2019

An Ensemble Method for Interval-Censored Time-to-Event Data

Interval-censored data analysis is important in biomedical statistics for any type of time-to-event response where the time of response is not known exactly, but rather only known to occur between two assessment times. Many clinical trials and longitudinal studies generate interval-censored data; one common example occurs in medical studies that entail periodic follow-up. In this paper we propose a survival forest method for interval-censored data based on the conditional inference framework. We describe how this framework can be adapted to the situation of interval-censored data. We show that the tuning parameters have a non-negligible effect on the survival forest performance and guidance is provided on how to tune the parameters in a data-dependent way to improve the overall performance of the method. Using Monte Carlo simulations we find that the proposed survival forest is at least as effective as a survival tree method when the underlying model has a tree structure, performs similarly to an interval-censored Cox proportional hazards model fit when the true relationship is linear, and outperforms the survival tree method and Cox model when the true relationship is nonlinear. We illustrate the application of the method on a tooth emergence data set.

stat.ME↗

arXiv2017

Survival Trees for Interval-Censored Survival data

Interval-censored data, in which the event time is only known to lie in some time interval, arise commonly in practice; for example, in a medical study in which patients visit clinics or hospitals at pre-scheduled times, and the events of interest occur between visits. Such data are appropriately analyzed using methods that account for this uncertainty in event time measurement. In this paper we propose a survival tree method for interval-censored data based on the conditional inference framework. Using Monte Carlo simulations we find that the tree is effective in uncovering underlying tree structure, performs similarly to an interval-censored Cox proportional hazards model fit when the true relationship is linear, and performs at least as well as (and in the presence of right-censoring outperforms) the Cox model when the true relationship is not linear. Further, the interval-censored tree outperforms survival trees based on imputing the event time as an endpoint or the midpoint of the censoring interval. We illustrate the application of the method on tooth emergence data.

stat.ME↗

arXiv2016

Survival trees for left-truncated and right-censored data, with application to time-varying covariate data

Tree methods (recursive partitioning) are a popular class of nonparametric methods for analyzing data. One extension of the basic tree methodology is the survival tree, which applies recursive partitioning to censored survival data. There are several existing survival tree methods in the literature, which are mainly designed for right-censored data. We propose two new survival trees for left-truncated and right-censored (LTRC) data, which can be seen as a generalization of the traditional survival tree for right-censored data. Further, we show that such trees can be used to analyze survival data with time-varying covariates, essentially building a time-varying covariates survival tree. Implementation of the methods is easy, and simulations and real data analysis results show that the proposed methods work well for LTRC data and survival data with time-varying covariates, respectively.

stat.ME↗

arXiv2016

On the Sensitivity of the Lasso to the Number of Predictor Variables

The Lasso is a computationally efficient regression regularization procedure that can produce sparse estimators when the number of predictors (p) is large. Oracle inequalities provide probability loss bounds for the Lasso estimator at a deterministic choice of the regularization parameter. These bounds tend to zero if p is appropriately controlled, and are thus commonly cited as theoretical justification for the Lasso and its ability to handle high-dimensional settings. Unfortunately, in practice the regularization parameter is not selected to be a deterministic quantity, but is instead chosen using a random, data-dependent procedure. To address this shortcoming of previous theoretical work, we study the loss of the Lasso estimator when tuned optimally for prediction. Assuming orthonormal predictors and a sparse true model, we prove that the probability that the best possible predictive performance of the Lasso deteriorates as p increases is positive and can be arbitrarily close to one given a sufficiently high signal to noise ratio and sufficiently large p. We further demonstrate empirically that the amount of deterioration in performance can be far worse than the oracle inequalities suggest and provide a real data example where deterioration is observed.

stat.ML↗

arXiv2013

Efficiency for Regularization Parameter Selection in Penalized Likelihood Estimation of Misspecified Models

It has been shown that AIC-type criteria are asymptotically efficient selectors of the tuning parameter in non-concave penalized regression methods under the assumption that the population variance is known or that a consistent estimator is available. We relax this assumption to prove that AIC itself is asymptotically efficient and we study its performance in finite samples. In classical regression, it is known that AIC tends to select overly complex models when the dimension of the maximum candidate model is large relative to the sample size. Simulation studies suggest that AIC suffers from the same shortcomings when used in penalized regression. We therefore propose the use of the classical corrected AIC (AICc) as an alternative and prove that it maintains the desired asymptotic properties. To broaden our results, we further prove the efficiency of AIC for penalized likelihood methods in the context of generalized linear models with no dispersion parameter. Similar results exist in the literature but only for a restricted set of candidate models. By employing results from the classical literature on maximum-likelihood estimation in misspecified models, we are able to establish this result for a general set of candidate models. We use simulations to assess the performance of AIC and AICc, as well as that of other selectors, in finite samples for both SCAD-penalized and Lasso regressions and a real data example is considered.

stat.ML↗