Knowledge library

This notebook tests whether standard portfolio allocation can improve an every-bar NASDAQ-100 trading strategy that is already burdened by transaction costs. It selects predictions using validation performance on the declared cost-feasible universe, then…

This notebook applies position-level risk controls to leading ETF allocation combinations while keeping each underlying prediction, concentration, and allocator fixed. It compares stop-losses, trailing stops, and time exits with the original strategy,…

This notebook explains how to turn a released, cross-sectionally ranked US firm characteristic panel into a keyed feature matrix for a factor study. It groups inputs into declared families, combines characteristics and interactions without fitting…

This US equities feature study explains how to generate features from estimated models without allowing future data into earlier observations. Its estimation schedule uses a history burn-in, fits parameters only on data before each output block, then…

This notebook applies TSMixer to ETF sequences using one globally shared function across funds. Each example contains an individual fund's history and covariates; temporal and feature interactions are learned across the panel, but one fund's observations are…

This notebook describes a stochastic discount factor (SDF) model for pricing the cross-section of US firm returns. Instead of first estimating factors and then applying them, it directly learns firm-month weights under a no-arbitrage condition: discounted…

This notebook analyzes reconstructed NASDAQ limit order books to describe intraday spreads and top-of-book depth, then examine whether order-flow imbalance is associated with subsequent bucket returns. It expresses spreads in basis points to compare stocks…

This notebook presents a deployment bridge between an offline machine-learning pipeline and QuantConnect's LEAN trading engine. It selects a model run reproducibly from a registry, exports its holdout predictions as date-indexed JSON, and uses a small…

This notebook evaluates four news-derived signals—weighted surprise, average sentiment, sentiment change, and article coverage—against forward stock returns. It computes a daily cross-sectional Spearman information coefficient, summarizes its mean,…

This document describes fitting temporal convolutional networks to NASDAQ 100 minute level microstructure features. Causal convolutions prevent a prediction from using later observations, while dilation lets successive layers capture patterns over multiple…

This analysis checks whether daily options and share data can support a weekly S&P 500 strategy that ranks constituents by thirty day at the money implied volatility and buys the highest ranked shares. Options provide the signal, while the portfolio holds…

This analysis compares predictions from several model families trained on monthly US stock characteristics to forecast next-month returns. It focuses on cross-sectional information coefficient, which measures how well a model ranks stocks within each month.…

This notebook compares methods for discovering relationships among a panel of ETF returns: NOTEARS for contemporaneous linear directed acyclic graphs, VAR-LiNGAM for lagged and instantaneous structure, PCMCI for conditional-independence links, and Granger…

This notebook demonstrates tuning LightGBM hyperparameters with Optuna's TPE sampler, using cross-sectional information coefficient as the objective. It combines early stopping to choose the number of boosting rounds with a custom pruning callback that…

This study converts registered model predictions into comparable S&P 500 option strategies. On weekly decision dates it ranks predicted returns, filters to a liquid universe, and sells equally weighted at-the-money straddles on the highest-ranked symbols.…

This notebook brings together five latent-factor approaches for modeling the S&P 500 options case study’s equity return cross-section. PCA estimates common movements from returns alone; IPCA maps characteristics to exposures linearly; a conditional…

This notebook teaches how to decode NASDAQ TotalView-ITCH binary messages and store them as structured data for later market microstructure analysis. It explains message framing, fixed-width field layouts, big-endian values, timestamps measured from…

This notebook explains a supervised autoencoder for predicting the direction of future US equity returns across multiple horizons. Its encoder feeds a reconstruction decoder, an auxiliary classifier, and a main classifier. Joint training combines…

This notebook builds a portfolio allocator that places a Temporal Fusion Transformer-style variable-selection network before an LSTM encoder. The selection network embeds each input feature separately and assigns softmax weights, allowing the model to vary…

This document presents a cross-market inventory of model-based feature artifacts from nine case studies. It reads parquet schemas rather than loading their rows, excludes identifier columns, counts feature columns, and groups names by tokens associated with…

This document explains how to build and inspect a feature matrix for a cross-sectional ETF momentum and rotation hypothesis. It defines each feature’s lookback and information lag, then constructs trailing returns, risk-adjusted returns, volatility, trend…

This document reframes a five-session equity return prediction by sampling daily data on Fridays. The label remains a five-session return, but on the weekly grid it spans about one model step. The notebook compares direct regression, using a fixed lookback…

This document describes refitting the configuration selected by earlier validation stages on all eligible pre-2021 history, then generating predictions for the 2021 holdout. It derives the training interval from the declared evaluation window, label buffer,…

This notebook evaluates gradient-boosted trees on equity option analytics, where features such as implied volatility, skew, term structure, and variance risk premium encode market expectations. It asks whether a nonlinear model can combine those forecasts…