DT plot: Finding the Optimal Discrimination Threshold for Binary Classifier

Learn how to use one lesser-known yet a highly important diagnostic plot

A substantial part of typical machine learning tasks is related to training algorithms that assign input data with one of the two predefined labels. Such algorithms are generally known as binary classifiers. There are a number of ways to visualize the performance of such classifiers, and arguably the most common tools for that are receiver operating characteristic (ROC) and precision-recall (PR) …

binary classification-metrics discrimination model-performance plot plotly python threshold