scikit-learn-compatible time-series cross-validation: purging, embargo, combinatorial purged CV, and deflated Sharpe ratios
Honest Lopez de Prado-style validation harness: Deflated Sharpe + Combinatorial Purged CV + alpha/beta + bootstrap — kills false 'beta illusion' edges before risking capital.
Empirical asset-pricing research framework (Python) with formal overfitting control — Combinatorial Purged CV, Deflated Sharpe Ratio, Probability of Backtest Overfitting. ~20 pre-registered experiments on a survivorship-free, point-in-time S&P 500 universe, with an honest public kill-log.
Living technical reference for Disuza Quantitative — private quantitative research laboratory, Madrid, Spain. Architecture, anti-overfit methodology (CPCV / DSR / PBO per López de Prado), regulatory posture. Source code proprietary.
Independent, MIT-licensed Python implementation of the López de Prado causal factor-investing and falsification framework. Run a falsification battery, measure how much of a backtest edge survives search and selection, and falsify your edge before you risk capital.
[ESP] Master’s Coursework (mIAx): Bayesian networks for financial causal discovery: DAG inference on macro asset returns and reproduction of the factor mirage (López de Prado & Zoonekynd, 2025).
An honest-invalidation study of cross-sectional return prediction on Brazilian (B3) equities — López de Prado pipeline, CPCV, Deflated Sharpe. Not a trading system.
[ESP] Master’s Coursework (mIAx): Deep learning for S&P 500 return forecasting: 64-model grid (MLP/RNN/CNN/mixed) + López de Prado investigation track and a model-driven portfolio.
[ESP] Master’s Coursework (mIAx) - Quantitative Finance Studies: fixed income, equity microstructure, derivatives, risk management, backtesting, ML preprocessing (López de Prado), fund of funds.
To associate your repository with the lopez-de-prado topic, visit your repo's landing page and select "manage topics."