An Improvement of FP-Growth Mining Algorithm Using Linked list

Abstract

Frequent pattern mining such as association rules, clustering, and classification is one of the most central areas in the data mining research. One of the foremost processes in association rule mining is the discovering of the frequent pattern. To draw on all substantial frequent patterns from the sizable amount of transaction data, various algorithms have been proposed. The proposed research aims to mine frequent patterns from the sizable amount of transaction database by using linked list. In this method, first scanning the database, the count of frequent 1- itemsets is searched using the hash map and for next itemsets, it is stored in the linked list, second scanning the database. The frequent 2-itemsets is generated using hash table and so on. So, the proposed research needs only two scans and this proposed method requires shorter processing time and smaller memory space.