OR-Tree Based Fuzzy Associative Classification Approach and Its Application

CACA doesnpsilat have very high accuracy because it filters some classified characters in partition quantitative attribute into different intervals. We propose a new class based fuzzy associative classification approach (CFACA), use fuzzy sets to categorize a quantitative attribute, find the centers of the fuzzy sets by clustering, define the fuzzy support, fuzzy confidence, and membership degree based ordered rule tree (OR-tree). We use the CFACA algorithm to mine fuzzy associative rules on the data sets of mobile communications marketing. Experimental results showed that CFACA exhibits a good performance in accuracy over CACA, and will have broad applications in fuzzy associative classification problems.