PRIMITIVE·GitHub·Other
Causal Tree Decomposition
Standalone causal-tree probability engine — thesis to weighted confidence, zero deps
Install
npm i causal-tree-decompositionThe core probability engine behind SimpleFunctions' thesis evaluation, extracted and open-sourced. Give it a thesis (free-form text) and it returns a directed-acyclic tree of sub-claims with calibrated confidences, rolled up to a single root probability.
Runs offline (no API required), zero runtime dependencies. Use it as a reasoning scaffold, a Bayesian update engine, or a post-hoc explainer for LLM judgments.
Quickstart
import { decompose, rollUp } from 'causal-tree-decomposition'
const tree = decompose('Fed cuts rates 50bps by Oct 2026')
const p = rollUp(tree) // 0..1
Tags
primitivetypescriptprobabilitybayesianreasoning
Related
- PRIMITIVE·OtherPrediction Market Edge DetectorDetect mispricings in prediction markets — filter, rank, monitor
- PRIMITIVE·OtherPrediction Market Regime ClassifierDetect market regime states (crisis, risk-off, risk-on, complacent) in real time
- SDK·MITSimpleFunctions CLIPrediction market infrastructure CLI — 42 commands, MCP server, REST API