Diagnoses of Learning Disability Students by Support Vector Machines with Sequential Forward Selection Techniques

• Main Authors: Chao Wei Huang, 黃昭瑋
• Other Authors: Ping Feng Pai
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/14117261947839035382

碩士 === 國立暨南國際大學 === 資訊管理學系 === 98 === Abstract The identification or diagnosis of learning disability (LD) students has been a critical issue for a long time because of the implicit characteristics of LD. The LD diagnosis procedures usually include interpreting some standard statistical tests or checklist scores compared to the scores from non-LD students. The conduction of procedures usually takes around one year. Therefore, the aim of this study is to develop an efficient and effective model integrate sequential forward selection and support vector machines (SFSSVM) techniques to analyze data of LD students. The sequential forward selection approach was employed to select important condition attributes, and the support vector machines model was used for classifying LD students. In addition, the determination of SVM parameters influences the classification accuracy of SVM models a lot. Thus, artificial immune system algorithm (IA) was utilized to select the parameters of SVM models. To demonstrate the performance of the proposed model, two other approaches, namely the logistic regression model (LR) and back propagation neural networks (BPNN) were applied to cope with the same LD data. Empirical results revealed that the SFSSVM model outperform the other two techniques. Therefore, the SFSSVM model is an alternative for diagnosing LD students.