National Changhua University of Education Institutional Repository : Item 987654321/15991

English | 正體中文 | 简体中文 | Items with full text/Total items : 6507/11669
Visitors : 30397101 Online Users : 488

RC Version 3.2 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.

Scope Adv. Search

Login ‧ Upload ‧ Help ‧ About ‧ Administer

NCUEIR > College of Management > Department of Information Management > Periodical Articles > Item 987654321/15991

Please use this identifier to cite or link to this item: http://ir.ncue.edu.tw/ir/handle/987654321/15991

Title:	應用決策樹於學習障礙鑑定之評估 Application of Decision Tree Algorithm to the Identification of Students with Learning Disabilities
Authors:	蘇妍如;吳東光;孟瑛如
Contributors:	資訊管理學系
Keywords:	決策樹;學習障礙;特徵選取;機器學習 Decision tree;Learning disabilities;Feature selection;Machine learning
Date:	2007.09
Issue Date:	2013-04-22T07:38:20Z
Publisher:	中華大學資訊學院
Abstract:	在學齡階段中，學障學生的障礙特徵不像其他身心障礙類別的學生如此明顯，其診斷與鑑定也較為困難，以致於很多疑似的學障學生可能在最需要教育介入的階段無法發現其障礙，致無法接受特殊教育服務；同時十二年安置政策的開辦使得學障鑑定量大增，在診斷時需要更為謹慎。學障鑑定在教育與升學的需求下，變得非常重要。本文期望能夠利用機器學習技術中的決策樹進行資料探勘之分類功能。並嘗試於進行決策樹資料學習以建立分類模型之前，先實施資料的預先處理將眾多的資料屬性予以簡化，期望透過資料屬特徵的選取來達到提昇分類的效能。經過本研究的實驗處理並綜合分析其實驗數據成果，發現在經由特徵選取且透過決策樹的分類與找出規則後，可大幅提高預測率，可以說是有達到本研究所期望的成果。 The characteristics of school-age children with learning disabilities (LD) are not as obvious as students with other physical or psychological disabilities. As a result, the diagnose of students with LD has long been a very difficult issue that requires a lot of time and extensive manpower. Accordingly, there may be many potential LD students un-identified and unable to receive appropriate special education services. This objective of this study is to apply data mining technique to the raw data that are used in manual LD diagnosis process, and try to find explicit rules that could assist LD diagnosis personnel in the future. The machine learning algorithm we adopted is decision tree algorithm. In addition, various feature selection pre-processing algorithms are experimented on the data prior to the application of decision tree technique. Our experimental results show that, through some proper pre-processing procedure, the rules that are generated by decision tree technique do considerably improve the classification accuracy and thus may be helpful in practical diagnosis application.
Relation:	Journal of Information Technology and Applications, 2(2): 107-115
Appears in Collections:	[Department of Information Management] Periodical Articles

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	730	View/Open

Loading...