以縮減交易資料機制提昇探勘關聯規則演算法之效能

Improving the Efficiency of Mining Association Rules Algorithms by Transaction Reduction

陳垂呈(Chui-Cheng Chen)

《資訊管理展望》 12卷2期 (2010/12) Pp. 133-147

https://doi.org/10.29715/JIMCSA.201012.0002

摘要

本論文修改Apriori演算法對候選項目組掃瞄交易資料庫的方式，加入刪除未包含候選項目組之交易資料的概念，提出兩個有效率的演算法，稱之為efficiency_Apriori演算法及efficiency_MQA-1演算法，分別探勘關聯規則、及包含有項目數量的關聯規則。從實驗評估中顯示，efficiency_Apriori演算法及efficiency_MQA-1演算法可分別有效提昇Apriori演算法、及MQA-1演算法的執行效能。

關鍵字

資料探勘 ； 關聯規則 ； Apriori ； MQA-1

並列摘要

This paper modifies the approach of scanning the transaction database of the Apriori algorithm for candidate itemsets, and adds the idea of the deleting the transaction data which do not contain the candidate itemsets. Two algorithms, called efficiency_Apriori and efficiency_MQA-1, are proposed to mine association rules and association rules including the quantities of items, respectively. The experiments show that the efficiency_Apriori algorithm and the efficiency_MQA-1 algorithm can effectively improve the performance of the Apriori algorithm and the MQA-1 algorithm, respectively.

並列關鍵字

data mining ； association rules ； Apriori ； MQA-1