An Apriori-like algorithm for Extracting Fuzzy Association Rules between Keyphrases in Text Documents

In this paper we present an algorithm for extracting fuzzy association rules between weighted keyphrases in collections of text documents. First, we discuss some classical approaches to association rule extraction and then we show the fuzzy association rules algorithm. The proposed method integrates the fuzzy set concept and the apriori algorithm. The algorithm emphasizes the distinction between three important parameters: the support of a rule, its strength, and its confidence. It searches for rules containing different number of phrases and having confidence level and strength level above certain thresholds. The algorithm makes the distinction between a small number of occurrences with high support intersections and large number of occurrences with low support intersections. Finally we present results of initial experiments on real-world data that illustrate the usefulness of the proposed approach.