Selected Publications

Signed iterative Random Forests refine iRF to identify stable, predictive, high-order interaction rules. We propose several importance measures for evaluating interactions in terms of their stability and predictive accuracy. Using signed interactions and the proposed importance measures, we evaluate regulatory interactions in early stage Drosophila embryos.
arXiv preprint

We developed a predictive, stable, and interpretable tool: iterative Random Forests (iRF). iRF discovers high-order interactions among biomolecules at the same order of computational cost as Random Forests (RF). We demonstrate the efficacy of iRF by finding known and novel interactions among biomolecules, of up to 5th and 6th order, in two data examples in transcriptional and co-transcriptional regulation.
PNAS

Teaching

I will not be teaching for the Fall 2018 semester. I have previously taught:

  • STAT 2: Introduction to Statistics
  • STAT 133: Concepts in Computing with Data
  • STAT 215A: Statistical Models: Theory and Application

Contact