Which statement is correct?

Every tree classifier is a linear classifier.

Every linear classifier is a special case of a tree.

Neither of the above.

Both of the above.

What is a weak classifier? A classifier with an error rate...

slightly below 100%.

slightly above 50%.

slightly below 50%.

Suppose you are given a training data set in a given feature space, and a procedure that generates random decision trees on that feature space. You generate 100 trees at random, independently of each other. Each tree partitions the sample space into rectangular regions. Now you use the training data to assign class labels to each cell, for each tree, in the same way as the usual tree training algorithm (each cell is assigned the label of that class of which it contains the largest number of training points). Given a new data point x, you apply each tree to predict a label for x, resulting in 100 predicted labels. Since the trees are random, these labels can be regarded as random variables. If so, are these variables:

Stochastically independent?

Stochastically dependent?