PortfolioStudy
Helpers & Utilities · AiDotNet.Finance.Trading.Evaluation
One-call research entry point: given the FULL history, split it walk-forward (train on the past, hold out a
strictly-later tail with an embargo gap), run every experiment through
PortfolioExperimentRunner, and return the outcomes ranked by holdout Sharpe. This is the honest
study loop end-to-end — the seam a researcher calls to compare candidate objectives / leverage / frictions /
agents on data the agents never trained on.
Methods
| Method | Summary |
|---|---|
IReadOnlyList<PortfolioExperimentOutcome> Run(IReadOnlyList<double[]> assetPrices, IReadOnlyList<double[]> featureColumns, double trainFraction, int embargo, int windowSize, double initialCapital, IReadOnlyList<PortfolioExperiment> experiments, Func<int, int, IPortfolioAgent<T>> agentFactory, int trainEpisodes) |