Package: kronxNBC 0.1.1

Overview | Step 1 — Feature Engineering | Rolling Volatility | Drawdown | Downside Semi-deviation (Transition Stress Proxy) | Residence Pressure | Ruin Proxy | Attach Regime Labels and Export | Step 2 — Why Random Sampling, Not a Chronological Split | Step 3 — Fitting the Student-t Naive Bayes Classifier | Step 4 — Inspecting the Fitted Parameters | Parameter Tables | Coefficient Data Frame | Density Plots | Step 5 — Out-of-Sample Evaluation | Predictions | Confusion Matrix | COR Stress Alert | Step 6 — Interpreting the $\nu$ Parameter | What low $\nu$ means | Validation of the heavy-tail hypothesis | Appendix — Session Info

Abstract | Introduction | Statistical Methodology | The Naive Bayes Framework | The Scaled Student-$t$ Distribution | Profile Maximum Likelihood for $\nu$ | Numerical Stability During Scoring | Package Overview | Installation | Function Reference | student_t_naive_bayes() | predict() | tables() | coef() | plot() | Complete Worked Example | Synthetic Dataset | Fitting the Model | Inspecting the Fitted Parameters | Parameter Tables | Coefficient Data Frame | Interpreting the $\nu$ Parameter | Density Visualisation | Out-of-Sample Evaluation | Class Predictions | Confusion Matrix | Posterior Probabilities | COR Stress Alert Signal | Comparison: Student-$t$ vs Gaussian Naive Bayes | Setup | Gaussian NBC via naivebayes | Head-to-Head Accuracy | Confusion Matrices | Stress-Class Recall | Tail Log-Density Comparison | Advanced Usage | Custom Prior Probabilities | Tuning the $\nu$ Grid | Accessing Raw Parameters | Predicting with a Subset of Features | The Clock of Regimes (COR) Framework | Data Requirements | Feature Engineering | Train/Test Split Strategy | Interpreting Empirical $\nu$ Estimates | Summary | Appendix: Session Information | References