README.md

In yuhuilin619/prism: Partially Recursively Induced Structured Moderation

prism:

Partially Recursively Induced Structured Moderation

Description:

We formulate a new highly interpretable tree-based modeling framework that allows for automatic discovery of potentially non-linear hierarchical interactions between health determinants at multiple levels (individual factors and contextual factors) and differences in survival outcomes between groups of interest. Inferential, predictive and visualization strategies are also developed. Finally, a new variable importance measurement is derived, so a local vimp (Lvimp), which helps to understand the individual level variables that are driving each terminal node.

License:

prism is open source / free software, licensed under the GPLv3.

Requirements:

prism requires R-3.4.0

Installation and usage:

install.packages("devtools") library("devtools") devtools::install_github("yuhuilin619/prism") library("prism")

References:

J.Sunil Rao, Erin Kobetz, Huilin Yu, Jordan Baeker-Bispo and Zinzi Bailey (2018) "Partially Recursively Induced Structured Moderation (PRISM) for better understanding the complexity of racial differences in endometrial cancer survival".

yuhuilin619/prism documentation built on July 18, 2019, 5:41 p.m.