Software | Dhrubajyoti Ghosh
I develop research software to make methodological work reproducible and usable beyond the original papers.
PFCI
Penalized FCI for sparse causal structure
learning.
R package for high-dimensional causal discovery with latent
confounding.
- CRAN
- GitHub
- Associated work: Penalized FCI for Causal Structure Learning in a Sparse DAG for Biomarker Discovery in Parkinson’s Disease.
ivgls
Network-aware IV regression with graph-fused
lasso.
R package for causal node discovery and instrumental-variable estimation
with graph-structured exposures.
LRST
Longitudinal Rank-Sum Test.
R software implementing rank-based procedures for multivariate
longitudinal studies, including the core LRST framework and related
design procedures.
Polyspectra
R package for polyspectra, polyspectral means, and asymptotic variance estimation.
QuadraticPredictionR
R package for quadratic forecasting and higher-order time-series prediction.
Software philosophy. I aim to accompany new methodology with reproducible code, examples, and reusable implementations whenever possible.