Back to browse
GitHub Repository

High-performance causal discovery for time series — VarLiNGAM, DirectLiNGAM, RCD in Rust. 14-50x faster than Python lingam.

4 starsRust

VarLiNGAM-rs / Causal discovery in Rust, 50x faster than Python

by edyos·Feb 20, 2026·1 point·1 comment

AI Analysis

●●SolidWizardryNiche Gem

14-50x faster Rust port of lingam, but causal discovery is a niche research domain.

Strengths
  • Native Python bindings via PyO3 with scikit-learn-style API reduces friction for adoption
  • Complete implementation of six algorithms (VarLiNGAM, DirectLiNGAM, RCD, FastICA, Scalable, Bootstrap) with validation framework
  • Solves a real pain: production-scale causal discovery bottleneck for multi-hundred-variable datasets
Weaknesses
  • Causal discovery is a specialized research domain—small addressable audience outside academia
  • No published benchmarks or reproducible comparisons against lingam on standardized datasets
Target Audience

Data scientists, ML engineers, and researchers working with time series causal inference

Similar To

python-lingam · DoWhy · pgmpy

Similar Projects

Data●●Solid

New Causal Impact Library

Rust-powered CausalImpact port that's 10-30x faster than the R original.

Niche GemBig Brain
djwjjtw
922mo ago