Cut at 0.4 instead of 0.5 and return [accuracy, recall], rounded to 4 places

Machine Learning · from Thresholds, ROC and PR

predict_probathresholdroc_curveroc_auc_scoreprecision_recall_curveaverage_precision_score

Putting the kettle on…

Starting up…

Exercise 2

Cut at 0.4 instead of 0.5 and return [accuracy, recall], rounded to 4 places. Compare the accuracy to the 0.7956 you already know.

your answer
stuck? the lesson behind it
Thresholds, ROC and PR
predict() hides a choice you should be making yourself