Roy229/wandb_filesystem_huggingface_terminal_6928-model-registry-1787472057x13238

Churn Prediction Model Registry

This repository tracks candidate models evaluated for the customer churn prediction use case.

Experiment Review Summary

Display Name	Model Type	Val Accuracy	Val Loss	Decision
churn_xgb_base	xgboost	0.712	0.421	REJECT
churn_xgb_tuned	xgboost	0.823	0.312	REJECT
churn_rf_base	random_forest	0.684	0.518	REJECT
churn_rf_tuned	random_forest	0.791	0.293	REJECT
churn_lgbm_tuned	lightgbm	0.856	0.274	PROMOTE
churn_cat_underfit	catboost	0.579	0.634	REJECT
churn_xgb_shallow	xgboost	0.801	0.398	REJECT

Promoted model: churn_lgbm_tuned (val_accuracy=0.856, val_loss=0.274)

Model Card

The model card will be updated after each experiment review cycle.