The Irrelevant Values Problem of Decision Tree for Improving a Glass Sputtering Process

機構典藏 > College of Engineering > Graduate Institute & Department of Computer Science and Information Engineering > Journal Article > Item 987654321/54953

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/54953

Title:	The Irrelevant Values Problem of Decision Tree for Improving a Glass Sputtering Process
Authors:	Chiang, Ding-An;Wang, Cheng-Tzu;Wang, Yi-Hsin;Chen, Chun-Chi
Contributors:	淡江大學資訊工程學系
Keywords:	Data Mining;Decision Tree;The Irrelevant Values Problem;Glass Sputtering Process;Yield Analysis
Date:	2010-12-01
Issue Date:	2011-08-01 15:30:09 (UTC+8)
Publisher:	新北市：淡江大學
Abstract:	In this paper, we use decision tree to establish a yield improvement model for glass sputtering process; however, the tree may have irrelevant values problem. In other words, when the tree is represented by a set of rules, not only comprehensibility of the resultant rules will be detracted but also critical factors of the manufacturing process cannot be effectively identified. From the performance issue and practical issue, we have to remove irrelevant conditions from the rules; otherwise, a domain expert is needed to review the decision tree. In this paper, we use a very simple example to demonstrate this point of view. Moreover, to identify and remove irrelevant conditions from the rules, we also revise Chiang's previous algorithm such that the modified algorithm can deal not only discrete data but also quantitative data.
Relation:	Tamkang Journal of Science and Engineering=淡江理工學刊 13(4), pp.413-422
DOI:	10.6180/jase.2010.13.4.07
Appears in Collections:	[Graduate Institute & Department of Computer Science and Information Engineering] Journal Article

Files in This Item:

File	Description	Size	Format
1560-6686_13(4)p413-422.pdf		868Kb	Adobe PDF	344	View/Open