Asymmetric Triangular Fuzzy Sets for Classification Models

摘要

Decision trees have already proved to be important in solving classification problems in various fields of application in the real world. The ID3 algorithm by Quinlan is one of the well-known methods to form a classification tree. Baldwin introduced probabilistic fuzzy decision trees in which fuzzy partitions were used to discretize continuous feature universes. Here, we have introduced a way of fuzzy partitioning in which we can have asymmetric triangular fuzzy sets for mass assignments approach to fuzzy logic. In this paper we have shown with example that the use of asymmetric and unevenly spaced triangular fuzzy sets will reduce the number of fuzzy sets and will also increase the efficiency of probabilistic fuzzy decision tree.