My GitHub: teazrq

• Dimension Reduction Forests (R package drforest)

  • • This package provides the statistical estimation methods for dimension reduction forests and local subspace variable importance.

    • Local subspace variable importance can be used for detecting influential variables in a personalized prediction/recommendation.

• Random Forests for Heterogeneous Treatment Effect Estimation with Multiple Responses (R package MOTE.RF)

  • • Designed for a setting in which the covariates X and multivariate response Y are both measured before and after the intervention

    • Applied to a microbiome study to estimate the dietary effect on multiple health outcomes

• A Nonnegative Matrix Factorization Tool Box (R package MatrixFact)

  • • This package implements a variety of matrix factorization tools, such as NMF, ONMF, semi-NMF, semi-orthogonal NMF, and some of their extensions to binary data.

    • Our method semi-orthogonal NMF can be applied to document-word matrix to extract meaningful information for analyzing text data.

• Orthogonality Constrained Optimization and Dimension Reduction (R package orthoDr)

  • • CRAN [link] and GitHub [link]. Examples

    • This package offers an optimization solver for orthogonality constrained problems: min f(X) s.t. X^TX = I by utilizing the algorithm developed by Wen & Yin (2013). We utilize this to solve a variety of semiparametric dimension reduction and personalized medicine problems.

• Reinforcement Learning Trees (R package RLT)

  • • On CRAN Task View: Machine Learning & Statistical Learning

    • You can find this example file which further explains the functions.

    • The survival analysis component is already implemented. However, some features are still under experiment.

• Recursively Imputed Survival Trees (RIST)

  • • Temporary R-code: RIST Example & ReadMe

    • R Package: will be integrated with RLT (still working on it...)