Association Rule Mining with Data Item including Independency based on Enhanced Confidence Factor

Along with the development of data collection and various storage technology, the large data of users activities in economy is stored. Extracting valuable information or knowledge regarding behavior of user from these data is becoming more and more important for marketing strategies of sales and commerce. Association rule mining is one of useful techniques in this application field and widely studied. But sometimes too many rules that generated by association rule mining usually caused the wrong decisions made by manager, parts of generated rules are meaningful and useful, but other generated rules are unnecessary for manager to make the right decisionsIn this paper, in order to extract useful rules efficiently, we proposed a new framework of association rule mining based on enhanced confidence factor. Thus, the certainty factor was introduced to identify different situations and analysis the accuracy of association rule mining respectively. We illustrate some merits of our proposed method by theoretical analysis. Our experiment results show that the sets of useful rules can be generated in a more efficient way by using our method, which means less and more accurate rules could be used to make the proper decisions by manager.