How it works...

A confusion matrix displays information about the actual and predicted classifications made by a model. The performance of such systems is evaluated with the help of data in the matrix.

The following table shows the confusion matrix for a two-class classifier:

 

The entries in the confusion matrix have the following meanings:

• TP is the number of correct predictions that an instance is positive
• FN is the number of incorrect predictions that an instance is negative
• FP is the number of incorrect predictions that an instance is positive
• TN is the number of correct predictions that an instance is negative