6. Evaluating Classifiers
# setup from mlwpy import * %matplotlib inline iris = datasets.load_iris() tts = skms.train_test_split(iris.data, iris.target, test_size=.33, random_state=21) (iris_train_ftrs, iris_test_ftrs, iris_train_tgt, iris_test_tgt) = tts
In the previous chapter, we discussed evaluation issues that pertain to both classifiers and regressors. Now, I’m going to turn our attention to evaluation techniques that are appropriate for classifiers. We’ll start by examining baseline models as a standard of comparison. We will then progress to different metrics that help identify different types of mistakes that classifiers make. We’ll also look at some graphical methods for evaluating and comparing ...