關聯式分類演算法結合規則優先權以改善分類之準確度

Improving the Performance of Associative Classification Algorithms with Rule Priorities

王務本(Wu-Pen Wang)

指導教授 ： 黃連進

淡江大學/工學院/資訊工程學系碩士在職專班/碩士(2011年)

https://doi.org/10.6846/TKU.2011.00175

摘要

雖然已有許多關聯式分類演算法被發表，但是都沒有將規則相依問題考慮進去。而規則相依問題會造成規則信賴度的改變甚至規則及類別的改變，進而影響到分類的結果，因此，要解決規則相依問題（找尋最佳規則執行順序）將是一個非常耗時的工作，本論文將提出Rule Priority演算法來對規則做排序，來達到較佳的執行順序，降低規則相依問題對分類結果產生的影響，進而改善最後分類的結果。因我們提出的演算法是一種時間多項式的演算法，所以可以很輕易的跟任何關聯式分類演算法結合。而在本論文中，我們將Lazy演算法加上規則優先權的概念，來與僅使用Lazy演算法的方式進行比較，而實驗結果也證明，規則相依性的確可以改善分類的精確度。

關鍵字

規則排序 ； 規則相依性 ； 關聯式法則 ； 關聯式分類演算法

並列摘要

Although different associative classification algorithms have been proposed, none of the available associative classification algorithms consider the rule dependence problem that directly influences the classification accuracy of associative classification algorithms. Since the finding of the optimal execution order of class association rules (CARs) is a combinational problem, instead of finding the optimal execution order of CARs, in this paper we propose polynomial time algorithms to re-rank the execution order of CARs by rules’ priority. This reduces the influence of rule dependency problems. Consequently, the performance (the classification accuracy and recall rate) of the associative classification algorithms can be improved. The experimental results show that using LAZY with our method can get better classification results than that of the LAZY association classifier without considering the rule dependence problem.

並列關鍵字

Associative Classification Algorithms ； Association Rule ； Ranking ； Rule Dependency

參考文獻

[1] Alipio M. Jorge, Paulo J. Azevedo, “An Experiment With Association Rules And Classification: Post-Bagging And Conviction”, Lecture Notes In Computer Science, vol 3735, Oct 2005, Pp.137-149, 2005

[2] Bingheng Yan, Depei Qian, “Building A Simple And Effective Text Categorization System Using Relative Importance In Category”, Proceedings Of The Third International Conference On Natural Computation (Icnc 2007), vol 01, Pp. 108-114, 2007

[4] Hamill Karen A. And Zamora Antonio, “The Use Of Titles For Automatic Document Classification,” Jasis, vol31, Pp. 396-402,1980

[5] Hisham Al-Mubaid, Syed A. Umair, “A New Text Categorization Technique Using Distributional Clustering And Learning Logic,” Ieee Transactions On Knowledge And Data Engineering, vol. 18, No. 9 Pp. 1156-1165, 2006

[7] F. THABTAH, “A review of associative classification mining,” Knowl. Eng. Rev., vol. 22, pp. 37-65, 2007