Public copy of my long-only stock ranking and paper-trading research system: XGBoost over 264 names, LLM thesis overlay, nightly pipeline, Alpaca paper account. No edge shown yet; see README.
-
Updated
Oct 4, 2026 - Python
Public copy of my long-only stock ranking and paper-trading research system: XGBoost over 264 names, LLM thesis overlay, nightly pipeline, Alpaca paper account. No edge shown yet; see README.
ML-based long-short factor investing pipeline for Taiwan stocks using Size, BM & Momentum factors — OLS, RF, NN, XGBoost, LSTM
Diffusion-flow generative model for alpha factor discovery — learns P(α) ∝ R(α) over expression space for diverse, reward-proportional alpha portfolios
Cross-sectional equity alpha research & backtesting engine: ~10 point-in-time factors on the S&P 500 with IC/ICIR analysis, decile long-short portfolios, walk-forward IC-weighted composite, transaction costs, and strict no-lookahead tests.
面向 B站网页游戏《FX 简单!》的量化交易程序:自适应多因子策略 + 反未来函数自检 + 可观测决策雷达
Tiny dependency-free calculator: compute the Pearson & Spearman rank Information Coefficient of a trading signal from predictions vs realized returns. By microalphas.com
PnL attribution and alpha decay framework for FinBERT sentiment signals — Information Coefficient decay curves, Spearman rank correlation, statistical significance bands, rolling signal quality monitoring.
Tiny dependency-free calculator: fit an exponential decay to a trading signal's Information Coefficient and read off its half-life + rebalancing cadence. By microalphas.com
Paper-only factor research: lineage, IC gates, point-in-time data, shadow-ledger settlement
Cross-sectional multi-factor stock selection — momentum/volatility/reversal factors, Information Coefficient evaluation, quantile portfolios, and composite alpha.
LLM-driven alpha-factor mining with decay-resistant regularization, evaluated against WRDS CRSP equity data.
Count Each Idea Once: robust composites from clusters of correlated alpha-signal variants (Berkeley MFE x Ultramarin, 2026). Package, tests, report and deck.
IC-Gated Deployment Framework (ICGDF): a two-stage statistical gate for cross-sectional equity ML. Mean IC -0.0005, gate closed in 0/12 folds, ECE < 0.025 despite zero discrimination. SSRN preprint 6742700.
To associate your repository with the information-coefficient topic, visit your repo's landing page and select "manage topics."