Domain-agnostic causal inference engine for time series data.
Implements temporal alignment, Pearson/Spearman correlation, Bradford Hill scoring, Predictor Impact Score (PIS), optimal value analysis, and effect size calculation.
Paper: dFDA Specification — Decentralized FDA framework for automated causal inference from personal health data.
Source: QMD
import { alignTimeSeries, calculateCorrelation, calculatePIS } from '@optimitron/optimizer';
const pairs = alignTimeSeries(predictor, outcome, { onsetDelay: 7, durationOfAction: 14 });
const correlation = calculateCorrelation(pairs);
const pis = calculatePIS(pairs, correlation);176 unit tests covering all algorithms.
pnpm test --filter @optimitron/optimizer