C4.5 algorithm

C4.5 is an algorithm that is used for decision making. It is used in decision trees and is the successor of ID3 algorithm.

Besides the well known CART's and CHAIDs C4.5 is becoming increasingly important. He is now already used by different software packages.

Basically, this algorithm behaves very similar to the CART algorithm. The main difference is that in C4.5 must be no binary division, but any number of branches can be inserted. The tree is wider. It is usually less deep than the corresponding CART tree. For this to be less important after the first classification, the following splittings.

Another difference is the so-called pruning, the stub of a tree. CART produces some subtrees and tests them with new, not previously classified data. C4.5 prunes the tree, however, without observing the given data base.

See also: ID3, machine learning, TDIDT

  • Classification methods