Add explainers/class-imbalance.md.
Why skewed positive/negative ratios wreck naive accuracy and disproportionately hurt minority subgroups, and how common fixes (oversampling, undersampling, SMOTE, class weights) can either help or introduce new bias. Pairs naturally with the Precision-Recall Curve request (#92).
Suggested structure (match roc-curve-auc.md): one-sentence definition → why it matters for fairness → core concept → concrete example tied to a repo audit → runnable detection code → limitations → related concepts/projects/further reading.
Related: roc-curve-auc, sampling-bias, false-positives-vs-false-negatives.
See CONTRIBUTING.md: add explainers/<slug>.md plus a one-line entry in assets/explainers-data.json; the build script generates the page.
Freeze-safe: explainers are fully open during the paper freeze. If you quote any Fair Code benchmark result, use the frozen numbers in paper/results-frozen/ (never re-run your own) — see CLAUDE.md.
Add
explainers/class-imbalance.md.Why skewed positive/negative ratios wreck naive accuracy and disproportionately hurt minority subgroups, and how common fixes (oversampling, undersampling, SMOTE, class weights) can either help or introduce new bias. Pairs naturally with the Precision-Recall Curve request (#92).
Suggested structure (match roc-curve-auc.md): one-sentence definition → why it matters for fairness → core concept → concrete example tied to a repo audit → runnable detection code → limitations → related concepts/projects/further reading.
Related: roc-curve-auc, sampling-bias, false-positives-vs-false-negatives.
See CONTRIBUTING.md: add
explainers/<slug>.mdplus a one-line entry inassets/explainers-data.json; the build script generates the page.