A Strategic Study of Mining Fuzzy Association Rules Using Fuzzy Multiple Correlation Measues

Two different data variables may behave very similarly. Correlation is the problem of determining how much alike the two variables actually are and association rules are used just to show the relationships between data items. Mining fuzzy association rules is the job of finding the fuzzy item-sets which frequently occur together in large fuzzy data set, where the presence of one fuzzy item-set in a record does not necessarily imply the presence of the other one in the same record. In this paper a new method of discovering fuzzy association rules using fuzzy correlation rules is proposed, because the fuzzy support and confidence measures are insufficient at filtering out uninteresting fuzzy correlation rules. To tackle this weakness, a fuzzy correlation measure for fuzzy numbers, is used to augment the fuzzy support-confidence framework for fuzzy association rules. We have extended the Apriori algorithm to fuzzy multiple correlation analysis, which is the new approach presented in this paper comparing to most of the previous works. A practical study over the academic behaviour of a particular school is done and some valuable suggestions are given, based on the results obtained.