Evaluating models

Whenever we obtain a result, it is is only as accurate as the model that represents the real problem. It is, therefore, extremely important to understand which methods can be used to evaluate the performance of our models.

When dealing with classification models we can use the following methods.

Analyzing classification accuracy

This is the ratio of the number of correct predictions (CP) to the total number of samples:

Here, CP is the number of accurate or correct predictions, and TP is the total count of all the predictions that have been made.

Building the confusion matrix

...