Intelligent techniques for temporal sequence pattern mining using association rules

Abstract

In todays internet era a vast amount of knowledge is stored in the web and database Due to the availability of huge knowledge repositories getting the relevant information is a challenging task and hence it must be mined and extracted Association Rule Mining is one approach for extracting useful knowledge from datasets or database which includes frequent patterns and association rules between the items or attributes of a dataset with varying levels of strength In this thesis an intelligent architecture for effective Frequent Itemset Mining and Association Rule Mining has been proposed and implemented in order to provide effective mining for both temporal database and conventional database The proposed architecture consists of six different components namely User interface Data Set Rule Base Decision Manager Frequent Itemset Generator component and User Profile Manager model which uses the User Preference Database for fixing the support threshold The Frequent Itemset Generator provides frequent itemsets using the standard data structure component Temporal Frequent Pattern Tree algorithm and Hashing with Quadratic Probing for effective data mining These mining methods are used to find frequent itemsets from temporal database as well as conventional transaction database respectively