FHAR: A New Text Association Rule Algorithm Based on Concept Vector and Its Application

摘要

A novel text association rule approach FHAR algorithm is presented. To overcome the defect of traditional keywords which does not take into account the semantic relation between keywords, FHAR algorithm in the paper is based on concept vector. The density of semantic field and the weight of meaning are used to determine the concept of the keywords, which not only adds the texts semantic, but also reduces vector dimensions, FHAR algorithm adopts improved HASH table for efficient large item set generation. The stored address of item sets is determined by a new hash function. Based on the new hash table, tree structure is constructed. When FHAR algorithm is applied to text mining, the text association rule is derived. Experiments show FHAR algorithm possesses higher efficiency and accuracy than Apriori algorithm.