4.1 Basic process
The purpose of decision tree is to produce a tree with strong generalization ability-use the test set to detect generalization ability
As shown in Figure 4.1, from the root node of the tree to the leaf nodes (that is, the discrimination result), there are generally several intermediate nodes, and each intermediate node corresponds to an attribute test, such as the color attribute in the figure. , Root attribute, knocking attribute. The root node contains the complete set of samples, and each time it passes through an intermediate node, it is divided into sub-nodes based on the result of the intermediate node's attribute test.
The basic process follows the **"divide and conquer" **divide and conquer strategy.
Three situations lead to recursive returns:
(1) The samples contained in the current node all belong to the same category and do not need to be divided; (the same as Y)
(2) The current attribute set is Empty y