Use Case Code Quality ML - GuitarAlchemist/ix GitHub Wiki
Navigation: Home | ML-Pipelines | Architecture | Governance
ML-powered code analysis: ix-code metrics → Random Forest classification → time series trend tracking.
ix-code analyzes source files for cyclomatic/cognitive complexity, Halstead metrics, SLOC, maintainability index. Produces 20-feature Array1<f64> vectors.
Random Forest trained on labeled (acceptable/refactor) code samples. Cross-validate with StratifiedKFold, evaluate with ConfusionMatrix + ROC/AUC.
Weekly maintainability index tracked with rolling_mean and ewma. Alert when rolling average drops below threshold (60).
lag_features_with_stats → LinearRegression with temporal_split. Forecast future maintainability from recent trends.
| Algorithm | ix Crate | Role |
|---|---|---|
| Code Analysis | ix-code | Extract complexity metrics |
| Random Forest | ix-ensemble | Refactoring classifier |
| Cross-validation | ix-supervised | Model evaluation |
| Confusion Matrix + AUC | ix-supervised | Quality metrics |
| Rolling Mean / EWMA | ix-signal | Trend monitoring |
| Lag Features | ix-signal | Time series → supervised |
| Linear Regression | ix-supervised | Debt prediction |