Look-ahead based fuzzy decision tree induction

Decision tree induction is typically based on a top-do,vn greedy algorithm that makes locally optimal decisions at each node. Due to the greedy and lo cal nature of the decisions made at each node, there is considerable possib ility of instances at the node being split along branches such that instanc es along some or all of the branches require a large number of additional n odes for classification. In this paper, we present a computationally effici ent say of incorporating look-ahead into fuzzy decision tree induction. Our algorithm is based on establishing the decision at each internal node by j ointly optimizing the node splitting criterion (information gain or gain ra tio) and the classifiability of instances along each branch of the node. Si mulations results confirm that the use of the proposed look-ahead method le ads to smaller decision trees and as a consequence better test performance.