PortfolioExperimentRunner
Helpers & Utilities ยท AiDotNet.Finance.Trading.Evaluation
Runs a set of reward experiments: for each, train a FRESH agent on the training split, then evaluate it and the equal-weight baseline on the UNTOUCHED holdout split, and rank by holdout Sharpe. This is the honest-eval harness for the trading-agent research โ chronological train/holdout, vs-baseline, ranked on risk-adjusted holdout performance, never training loss.
Methods
| Method | Summary |
|---|---|
IReadOnlyList<PortfolioExperimentOutcome> Run(IReadOnlyList<double[]> trainAssetPrices, IReadOnlyList<double[]> trainFeatureColumns, IReadOnlyList<double[]> holdoutAssetPrices, IReadOnlyList<double[]> holdoutFeatureColumns, int windowSize, double initialCapital, IReadOnlyList<PortfolioExperiment> experiments, Func<int, int, IPortfolioAgent<T>> agentFactory, int trainEpisodes) |