多層次排序關聯分類器

Associative Classifier with Multi-Ranking

楊正奇(Jeng-Chi Yang)

指導教授 ： 蔣定安

淡江大學/工學院/資訊工程學系碩士班/碩士(2009年)

https://doi.org/10.6846/TKU.2009.00608

摘要

關聯法則是經常被使用在資料探勘的研究技術之一，進而利用關聯規則結合成關聯分類器，預測未知分類資料的類別。

關聯規則執行分類前，會依照演算法所定義的執行順序做排序。一般而言，當規則經過排序後，規則與規則之間的執行先後順序將不再改變。實際上關聯規則執行時，當排序較高的執行後，排序較低未執行的規則在剩餘未分類資料中，可能擁有較原先更高的信賴度，或更低的信賴度，規則間的執行先後順序與重要性可能會有所不同。

因此，本論文對此種情況，利用關聯法則提出多層次排序(Multi-Ranking)分類器，定義關聯分類器的規則執行順序及規則修剪的方法。由實驗結果顯示，多層次排序分類器在預測未知分類能有好的準確率表現與執行效率。

關鍵字

關聯分類器 ； 關聯法則 ； 執行順序 ； 規則修剪 ； 資料探勘

並列摘要

Association rule is one of the .adopted techniques frequently in data mining, then integrating the association rules into a associative classifier for predicting that data are not classified.

The association rules will be sorted by the algorithm’s definition before executing the association rules. In general, between the rule and the rule execution order no longer will change successively when the association rules are sorted. Actually, after executing higher rank , the lower ranking and unexecuted rules will have different confidence from initial confidence in the remaining data. And the rules’ execution order and importance will be difference.

Therefore, we propose a new classifier named Multi-Ranking classifier in view of the situation, defining the rules of the associative classifier execution orders. Moreover, Multi-Ranking classifier have good accuracy and execution performance in the experiment.

並列關鍵字

Associative Classifier ； Association Rule ； Execution Order ； Rule Pruning ； Data Mining

參考文獻

[2] Quinlan, J. R., C4.5: Programs for Machine Learning, Morgan Kaufmann, San Mateo, CA, 1993.

[4] G. V. Kass, “An Exploratory Technique for Investigating Large Quantities of Categorical Data”, Applied Statistics, pp.119-127, 1980.

[6] B. Liu, Y. Ma, and C. K. Wong, Improving an Association Rule Based Classifier., PKDD-2000, 2000, pp. 504-509.

[7] G. Dong, X. Zhang, L. Wong, and, J. Li, CAEP: Classification by aggregating emerging patterns., DS-99 (LNCS 1721), Japan, Dec. 1999.

[8] W. Li, J. Han, and J. Pei, CMAR: Accurate and efficient classification based on multiple class-association rules., ICDM-01, San Jose, CA, Nov. 2001, pp. 369-376.