Research Of Recommendation Methods Based On High Utility Pattern

In the rapid development of modern society, recommendation system has been widely used in many fields, and has good research and development space, so that it has become a hot spot in the academic field. In the related technology of recommendation system, association rules can be said to be one of the most commonly used method. But the association rules treat all items equally. The result is that it can not handle the data in which the items has differences. However, the high utility recommendation method solutes the problem well. Using utility values record the differences between items. Just because of this, the high utility recommendation method can apply in analysis of profits, revenue ratio and other fields. And this is of great significance.Mining high utility itemsets is becoming hot research topic in data mining owing to its ability to reflect users’bias and make up for the shortcoming of measuring itemsets only by supports. To meet the needs of larger and larger data size, a parallel algorithm, called PHUI-Mine (Parallel High Utility Itemset Mine), for mining high utility itemsets is proposed. Firstly, a tree structure, called the DHUI-tree (Dynamic High Utility Itemset Tree), is introduced to capture information of high utility itemsets. Meanwhile, the DHUI-tree construction method is described, and the dynamic pruning strategy of DHUI-tree is discussed. Then, the parallel algorithm is presented. Experimental results show that PHUI-Mine algorithm is efficient and consumes low storage cost.In the Consumption system of Chinese herbal medicine, the algorithm of high utility itemset mining is applied and herb price of every herb is considered as utility value in the high utility mining. we analyse and get the most valuable high utility frequent itemsets which used to direct implementation of process and recommend to the customers. According to the analysis of the herb information in recent years, we calculate the coverage and precision. Through the real date, we conclude that the method is very valuable...