Try 40K dataset (4x data)

[dberenbaum]

We are trying here a large dataset, since the smaller one looks unstable

[github-actions]

CML Report

Plots

Feature Importance: main

Feature Importance: workspace

Metrics and Params

main → workspace

Path	Metric	main	workspace	Change
eval/metrics.json	avg_prec.test	0.925	0.94496	0.019958
eval/metrics.json	avg_prec.train	0.97437	0.97723	0.002859
eval/metrics.json	roc_auc.test	0.94602	0.96191	0.015888
eval/metrics.json	roc_auc.train	0.98667	0.98737	0.00069973

Path	Param	main	workspace	Change
params.yaml	featurize.max_features	200	500	300