probe: Sparse high-dimensional linear regression with a partitioned empirical Bayes ECM algorithm

The probe package contains the R software tools to run the PaRtitiOned empirical Bayes Ecm (PROBE) algorithm. PROBE uses minimal prior assumptions on the parameters are used through the use of plug-in empirical Bayes estimates of hyperparameters. Efficient maximum (MAP) estimation is completed through a Parameter-Expanded Expectation-Conditional-Maximization (PX-ECM) algorithm. The PX-ECM results in a robust computationally efficient coordinate-wise optimization, which adjusts for the impact of other predictor variables. The completion of the E-step uses an approach motivated by the popular two-groups approach to multiple testing. The PROBE algorithm is applied to sparse high-dimensional linear regression, which can be completed using one-at-a-time or all-at-once type optimization. PROBE is a novel alternative to Markov chain Monte Carlo, empirical Bayes, and Variational Bayes approaches to fitting sparse linear models. See McLain et al. (2022) for more details.

Installation instructions for probe are available on GitHub. The probe package is available on CRAN with documentation for use.

lmmprobe: Sparse high-dimensional linear mixed modeling with a partitioned empirical Bayes ECM algorithm

lmmprobe is an R package extends the PROBE algorithm to sparse high-dimensional linear mixed-effect regression models. High-dimensional longitudinal data is increasingly used in a wide range of scientific studies. To properly account for dependence between longitudinal observations, statistical methods for high-dimensional linear mixed models (LMMs) have been developed. However, few packages implementing these high-dimensional LMMs are available in the statistical software R. Additionally, some packages suffer from scalability issues. This package implements Linear Mixed Modeling with PaRtitiOned empirical Bayes ECM (LMM-PROBE), an efficient and accurate Bayesian framework for high-dimensional LMMs. We use empirical Bayes estimators of hyperparameters for increased flexibility and an Expectation-Conditional-Minimization (ECM) algorithm for computationally efficient maximum a posteriori probability (MAP) estimation of parameters. The novelty of the approach lies in its partitioning and parameter expansion as well as its fast and scalable computation. See Zgodic et al. (2022) for more details.

Installation instructions for lmmprobe are available on GitHub.

wabh: False Discovery Rate Control for Lesion-Symptom Mapping with Heterogeneous data via Weighted P-values

This repository contains code to implement the Weighted Adaptive Benjamini Hochberg (WABH) procedure, which uses p-value weighting to improve the power of high-dimensional multiple testing. See Zhang et al (2023) for more details.

cdscr: Length-biased semicompeting risks models for cross-sectional data: An application to current duration of pregnancy attempt data

This repository contains programs for the analysis of current duration data that is subjected to semi-competing risks. The details of this model are discussed in McLain et al. (2021).

PHMM: Prediction intervals for penalized longitudinal models with multisource summary measures: An application to childhood malnutrition

This repository contains code to fit a Penalized Heterogeneous Mixture Model (PHMM). The details of the statistical methods can be seen in McLain et al. (2019). These programs estimate confidence and prediction intervals using an asymptotic estimate of the covariance matrix.

The methods in PHMM are used by UNICEF, WHO, and the World Bank to generate global health estimates for stunting and overweight. See mnf-sdg-stunting-overweight for the GitHub repository containing code and sample data for this project. For more, see the documentation from the Joint-Malnutrition-Estimates.