Mining Interesting Rare Items with Maximum Constraint Model Based on Tree Structure

Rare association rule mining provides useful information from large database. Traditional association mining techniques generate frequent rules based on frequent item sets with reference to user defined: minimum support threshold and minimum confidence threshold. It is known as support-confidence framework. As many of generated rules are of no use, further analysis is essential to find interesting Rules. Rare association rule contains Rare Items. Rare Association Rules represents unpredictable or unknown associations, so that it becomes more interesting than frequent association rule mining. The main goal of rare association rule mining is to discover relationships among set of items in a database that occurs uncommonly. We have proposed a Maximum Constraint based method for generating rare association rule with tree structure. Tentative results show that MCRP-Tree takes less time for rule generation compared to the existing algorithm as well as it finds more interesting rare items.