Classifier Evaluation

As mentioned in the section How does machine learning work?, the evaluation is one central part of every machine learning algorithm. It may seem trivial at first sight. Yet, deciding on the right measure is a crucial step. When we optimize our algorithm towards better performance, we will inevitably optimize the scores in our evaluation function.

We’ll get to know about more sophisticated evaluation functions in this course. But right now, let’s keep it simple. For example, a better way to evaluate the performance of a classifier is to look at the confusion matrix.