Knowledge library

This notebook turns 21-day forward ETF returns into an up-or-down target and evaluates L1- and L2-regularized logistic regression on chronological walk-forward folds. It scales features using each fold’s training data, then assesses predictions with…

This notebook configures Feast as a feature store for financial and model-based data, then compares its historical and point-in-time retrievals with a transparent, hand-built join. It declares entities, timestamped sources, feature views, and a time to live,…

This notebook compares gradient boosting objectives on monthly US stock returns while holding features, target, and walk-forward folds fixed. It contrasts squared error with absolute error and Huber loss, which reduces the influence of large residuals. The…

This module coordinates walk-forward fitting and evaluation for several latent factor models, including principal components, instrumented principal components, and neural autoencoders. It prepares panel data and folds, invokes model-specific runners, and…

This implementation describes a conditional autoencoder for asset pricing. A beta network maps each stock’s characteristics to latent factor loadings, while a linear factor network maps managed portfolio returns to factor returns. The model predicts a stock…

This reference describes the Financial Phrasebank, a labeled collection of financial news sentences used to train or evaluate natural language processing models. Human annotators assign positive, neutral, or negative sentiment labels. The corpus is available…

This analysis explains how to read cross-sectional information coefficients (ICs) for CME futures models and how they relate to later portfolio tests. It computes each date’s rank correlation between predicted and realized returns, then averages those…

This notebook compares BM25 keyword search, dense embeddings, reciprocal rank fusion (RRF), and cross-encoder reranking on passages from company 10-K filings. It builds each retrieval approach and assesses ranked results for exact, conceptual, and mixed…

This notebook compares LightGBM models on a US equities panel across three return horizons. Its grid varies tree leaf counts and loss functions, then scores each configuration at multiple tree-count checkpoints. The discussion explains how shallow trees can…

This notebook compares equal weight, inverse volatility, minimum variance, and hierarchical risk parity using the same ETF universe, forecasts, rebalance schedule, and backtest protocol. A rolling Ridge model ranks ETFs from momentum, moving-average…

This US equities case study distinguishes predicting future returns from claiming that momentum causes them. It treats a stock’s prior-year return excluding the latest month as the treatment, forward return as the outcome, and recent volatility, illiquidity…

This notebook develops a simulated market-making task in which a PPO agent chooses quote skew and spread width. Fill probability declines as quotes move farther from the mid price, while order imbalance affects both fills and subsequent price moves. A…

This teaching notebook compares target-label methods for supervised trading models using ETF data. It covers fixed-horizon returns and direction labels, rolling time-series percentiles, cross-sectional percentiles, triple-barrier labels with fixed or…

This shared modeling module supports quantitative research workflows by loading features, labels, temporal data, and configuration, then preparing walk-forward cross-validation folds. Its design emphasizes reproducibility and data lineage: hashes capture…

This notebook shows how to turn quarterly SEC 10-Q Management’s Discussion and Analysis text into two candidate equity signals. FinBERT scores text chunks for positive, neutral, or negative tone; chunk scores are aggregated into average sentiment and…

This notebook compares three ways to use market regimes in forecasting: a baseline model without a regime input, a single model given the filtered probability of a volatile state, and separate models trained within hard regime assignments. To avoid…

This notebook estimates whether buyer-initiated signed volume share is associated with future 15-minute returns after accounting for relative spread, recent realized volatility, and trailing monthly return. Double machine learning predicts the outcome and…

This notebook uses Double Machine Learning (DML) to estimate the adjusted effect of a continuous ETF momentum measure on subsequent returns, controlling for recent and longer-term volatility, market regime, and yield-curve slope. It compares unadjusted…

This notebook sets up an LSTM forecasting run for an FX-pairs case study. The model carries a hidden state through a lookback sequence, and the notebook resolves architecture and training settings through the shared configuration and study-planning tools. It…

This notebook examines how return forecasts for ETFs can express uncertainty, comparing Monte Carlo dropout with deep ensembles. Dropout remains active across repeated inference passes, while an ensemble measures disagreement among separately trained…

This notebook gathers registered double machine learning estimates from multiple trading case studies and compares their uncertainty and diagnostics. It loads current rows from each study’s registry, checks registry integrity and label uniqueness, makes…

This notebook fits a declared LSTM model for forecasting in an S&P 500 options study. The network receives chronological windows of each symbol’s history and uses gated recurrent state to represent information across sessions, rather than relying only on…

This notebook compares three ways to classify financial-news sentences as positive, neutral, or negative: TF-IDF word and phrase features with logistic regression, averaged static word vectors with a classifier, and a pretrained FinBERT sentiment model.…

This notebook evaluates accounting and price-based characteristics as standalone predictors of next-month US stock returns. It computes monthly cross-sectional rank correlations between each characteristic and the return label, then summarizes them across…