How to evaluate my Classification Model results

Songhao Wu
Towards Data Science
8 min readJun 14, 2020

--

Confusion Matrix, ROC-AUC, F1 Score, Log Loss

Source: photo by Annie Spratt from Unsplash

Different from the Regression model, when we assess classification model results, we should not only focus on prediction results but also need to consider the prediction probability and prediction ability of each class.

There are plenty of articles online about classification metrics selection and here I will…

--

--