Generates Alpha101/Alpha1921 football factors via ML-driven mining and RL-based evaluation workflows.
npx skills add https://github.com/Alsac/fminer --skill factor-mining
Please help me install this Agent Skill. Skill: factor-mining Source: https://github.com/Alsac/fminer/tree/main/.cursor/skills/factor-mining Command: npx skills add https://github.com/Alsac/fminer --skill factor-mining
This Skill provides an end-to-end platform for discovering, evaluating, and admitting alpha factors across multiple markets by combining large language models with reinforcement learning, reducing manual factor engineering work.
High-intent search queries and answers about installing and using this skill.
Alpha factor mining is automated by using LLM prompts to generate candidate formulas from Alpha101 and Alpha191 seeds, then evaluating them via GPU-accelerated backtesting and IC/ICIR analysis.
IC and ICIR analysis evaluates candidate factors by running GPU-accelerated computations on market data, checking backtesting performance and correlation, then admitting top factors into a managed repository.
Backtesting supports both A-share astock and crypto markets through configurable multi-market data pipelines that handle market-specific data loading, calibration, and reporting workflows.
A GPU is required for the factor discovery workflow because the evaluation stage uses GPU-accelerated computations to process IC/ICIR analysis and backtesting across multiple markets.
To generate a professional report for top factors, run the professional_report.py tool included in the scripts directory to produce a factor analysis tear sheet.
The best way to batch generate and test quantitative factors is running the batch_mine_parallel.py script, which orchestrates LLM-driven formula generation and RL-based evaluation.
VernonOYComputes factor quality via IC, ICIR, quintile, and robustness analyses.
HKUDSCompute IC/IR statistics and quantile backtest curves for factor selection.
DaddyElonMusk69Scale RL-driven LLM fine-tuning with Megatron-LM and SGLang rollouts.
0xZKnwEvaluate trading factor predictive validity via IC/IR analysis and quantile backtesting.
