Book Guide¶
The public Machine Learning for Trading, Third Edition companion shows
Diagnostic in longer research workflows. The links below point to notebooks
at book commit 2d6e8f95eeccaee66906245606471f570b5807e5.
All selected notebooks directly call ml4t.diagnostic; some also teach the
underlying method manually. The complete book workflows may require data and
optional packages beyond the synthetic quickstart.
Start with a Diagnostic guide for supported inputs and result checks, then open the book notebook for its surrounding research workflow. A call in a notebook does not make the entire book workflow a supported library API.
Validation and signal research¶
| Book notebook | Classification and use | Start in Diagnostic |
|---|---|---|
| Ch06: CV foundations | Direct CV splitter calls, plus manual fold construction | Cross-validation |
| Ch07: multiple testing | Direct statistical-test calls, plus method exposition | Statistical tests |
| Ch07: causal sanity checks | Direct Diagnostic checks in a broader causal review | Statistical tests |
| Ch08: feature selection | Direct feature-statistic and importance calls | Feature selection |
| Ch08: robustness and sensitivity | Direct Diagnostic signal checks in a wider workflow | Signal analysis |
| Ch08: event studies | Direct EventStudyAnalysis use | Evaluation workflows |
| Ch09: visual diagnostics | Direct feature-diagnostic calls | Feature diagnostics |
| Ch09: ARIMA features | Direct stationarity and autocorrelation checks in an ARIMA lesson | Feature diagnostics |
| Ch09: GARCH volatility | Direct volatility diagnostic in a GARCH lesson | Feature diagnostics |
| Ch09: Wasserstein regimes | Direct drift-distance call in a regime workflow | Feature diagnostics |
Backtests, portfolios, and trades¶
| Book notebook | Classification and use | Start in Diagnostic |
|---|---|---|
| Ch16: performance reporting | Direct backtest-bridge and reporting calls | Backtest tearsheets |
| Ch16: Sharpe inference | Direct Sharpe-inference calls, plus manual explanation | Statistical tests |
| Ch16: DSR validation | Direct deflated-Sharpe calls and validation examples | Statistical tests |
| Ch16: RAS protocol | Direct RAS adjustment calls | Statistical tests |
| Ch17: portfolio metrics | Direct PortfolioAnalysis use | Evaluation workflow |
| Ch17: mean-variance optimization | Direct portfolio diagnostics in an allocator lesson | Evaluation workflow |
| Ch17: robust optimization | Direct portfolio diagnostics in an allocator lesson | Evaluation workflow |
| Ch17: Kelly criterion | Direct portfolio diagnostics in a sizing lesson | Evaluation workflow |
| Ch17: hierarchical risk parity | Direct portfolio diagnostics in an allocator lesson | Evaluation workflow |
| Ch17: library comparison | Direct Diagnostic comparison calls | Evaluation workflow |
| Ch17: allocator comparison | Direct Diagnostic metrics in a broader allocator comparison | Evaluation workflow |
| Ch19: VaR and CVaR | Direct distribution diagnostic in a risk lesson | Feature diagnostics |
| Ch19: exit strategies | Direct BarrierAnalysis use | Evaluation workflows |
| Ch19: factor exposure | Direct factor-analysis calls | Evaluation workflows |
| Ch19: trade-SHAP diagnostics | Direct TradeAnalysis and TradeShapAnalyzer use | Trade analysis |
| Ch19: drift detection | Direct drift-diagnostic calls | Feature diagnostics |
The case-study evaluation notebooks show the same statistical checks across nine datasets.