A Classification Algorithm Based on an Association Rule of Multiple Frequent Item-Sets

Abstract

It is necessary to discrete datasets firstly if you want to data mining an association rule of datasets consisting of many categorical and numeric attributes by a traditional algorithm. However, in view of the versatility, the applications of the traditional algorithm are limited. This paper propose a new algorithm called ARMFI(Association Rule of Multiple Frequent Item-sets) which can data mining an Association Rule from datasets consisting of many categorical and numeric attributes directly and completely, and overcome disadvantage of the traditional algorithm. The result has been proofed that the ARMFI shows better performances than the traditional algorithm¿