High-Dimensional Adversarial Random Forests

This repository contains the code accompanying the paper "High-Dimensional Adversarial Random Forests". Before running any R script, set the working directory to the root of this repository.

The code is organized in the R-code directory, where scripts are numbered according to the analysis workflow.

Downstream Clustering Analysis

• 01-prepare-hember-data.R: Prepares the Hember dataset.
• 02-preprocess-hember-data.R: Preprocesses the Hember dataset.
• 03-hember-problem.R, 04-hember-algorithm.R, 05-hember-submission.R, 06-hember-results.R: Implement and evaluate the downstream clustering experiments using batchtools.
• 07-hember-pairwise-cor.R: Computes pairwise Pearson correlations between features.

Downstream Prediction Analysis

• 08-prepare-curated-tcga-data.R: Downloads and preprocesses the TCGA datasets.
• 09-tcga-problem.R, 10-tcga-algorithm.R, 11-tcga-submission.R, 12-tcga-results.R: Implement and evaluate the downstream prediction experiments.

Case Study: Alzheimer's Disease Brain Metabolomics

• 15-AD/: Contains all scripts for data preparation and analysis of the Alzheimer's disease brain metabolomics case study.
• 15-AD/01-ad-problem.R, 15-AD/02-ad-algorithm.R, 15-AD/03-ad-submission.R, 15-AD/04-ad-results.R: Implement and evaluate the downstream prediction analysis for the Alzheimer's disease dataset.