Enhancing the Performance of Mining High Utility Itemsets Based On Pattern Algorithm

Data Mining is the process of analyzing data from different perspectives and summarizing it into useful information. An association in data mining indicates a logical dependency between various attributes of an entity. Association rule mining (ARM) is the process of mining past data for association rules. ARM only find the frequency of itemsets, which will not provide large amount of profit. Utility mining focuses on discovering the itemsets with high sales profit. Here, utility mining is a measure of profitability of items to the users. The utility mining of itemsets is an important task in decision-making process of many applications such as website click streaming analysis, cross marketing in retail stores and in biomedical applications. The extraction of the high utility itemsets from a large database involves the creation of new candidate itemsets with high utility. This affects the performance of the mining process in terms of the execution time and the space requirement. In this paper, it is intended to develop an efficient algorithm for mining the high utility itemsets for reducing the candidate itemsets. Here, a data structure named pattern tree would be maintained to store the information about the high utility itemsets, so that the number of database scans can be reduced. .