Confusion Matrix in Three-class Classification Problems: A Step-by-Step Tutorial

Abstract

The confusion matrix is a specific table used in machine learning to describe and assess the performance of a classification model (e.g., an artificial neural network) for a set of test data whose actual distinguishing features are known. The confusion matrix for an n-class classification problem is square, with n rows and n columns. The rows represent the class actual samples (instances), which are the classifier inputs, and the columns represent the class predicted samples, which are the classifier outputs. Binary class classifiers have been presented in a previous paper, where in this paper, we are concerned with three-class classification performance measures. We also clarify the concept with numerical examples to make it close to the reader mind.