In Tutorial 5 you worked with the mushroom data set.

This data set contains an attribute stalk-root, which had some missing values, which were coded as ? in the data.

How many observations have a value of ? for the stalk-root attribute?

Models which generalise well capture the noise in the data in addition to the signal.

In a decision tree, attributes can be used more than once in segmentations and can be used with different split points.

Entropy is maximised when each class has equal probability of occurring.

Consider training a decision tree on the following dataset (with X1 and X2 representing 2 attributes, and Y the target variable):

Which one of the following splits at the root node gives the highest information gain?

Tree-structured models cannot give us estimates of the probability of customer churn.

In a classification tree, a non-leaf node is referred to as a "decision node" because it allows us to give a class prediction.

Which one of the following describes the rules represented in the decision tree below?

Entropy...

What has been the "muddiest point" for you thus far in the module?