Essential session storage is required to sign in. Optional analytics remain off unless you allow them.
Extend into a rigorous comparison. Requirements: build a scikit-learn PIPELINE (preprocessing + model) so transforms are fit per fold (leakage-safe); compare at least THREE algorithms (e.g. logistic regression, random forest, gradient boosting) using CROSS-VALIDATION; TUNE hyperparameters (grid or random search) on validation folds — never the test set; DIAGNOSE overfitting via the train-vs-validation gap and apply a remedy (regularization, a simpler model, or better features); and report results with a confusion matrix and the metric that matches the problem. Justify your final model choice with evidence, and explain why you tuned on validation rather than on test.
A Pipeline fitting preprocessing inside each CV fold; 3+ algorithms compared via cross-validation.
Hyperparameter tuning on validation (not test); overfitting diagnosed via the train/validation gap with a remedy applied.
A justified final-model choice with metrics + a confusion matrix; explains tuning-on-validation-not-test.
Use this README structure to make the work reviewable.
DevPath fetches a read-only, size-capped snapshot. Submitted code is never executed.
# Model comparison and tuning with a pipeline ## Goal Extend into a rigorous comparison. Requirements: build a scikit-learn PIPELINE (preprocessing + model) so transforms are fit per fold (leakage-safe); compare at least THREE algorithms (e.g. logistic regression, random forest, gradient boosting) using CROSS-VALIDATION; TUNE hyperparameters (grid or random search) on validation folds — never the test set; DIAGNOSE overfitting via the train-vs-validation gap and apply a remedy (regularization, a simpler model, or better features); and report results with a confusion matrix and the metric that matches the problem. Justify your final model choice with evidence, and explain why you tuned on validation rather than on test. ## Acceptance criteria - [ ] Leakage-safe pipeline: A Pipeline fitting preprocessing inside each CV fold; 3+ algorithms compared via cross-validation. - [ ] Tuning & diagnosis: Hyperparameter tuning on validation (not test); overfitting diagnosed via the train/validation gap with a remedy applied. - [ ] Evidence-based choice: A justified final-model choice with metrics + a confusion matrix; explains tuning-on-validation-not-test. ## Verification - [ ] Document installation and run commands. - [ ] Record automated checks and their exact commands. - [ ] Add representative output, screenshots, or a short demo where useful. - [ ] Confirm failure paths and known constraints. ## Decision log ### Decision title - Context: - Choice: - Alternatives considered: - Tradeoffs: ## Evidence - Link each rubric criterion to the file, test, or artifact that demonstrates it. ## Known limitations - List what is intentionally out of scope and what should be improved next.