Browsing by Author "Li, R. K."

Now showing 1 - 2 of 2

SVM-RFE Based Feature Selection and Taguchi Parameters Optimization for Multiclass SVM Classifier
(2014-01-01)
Huang, Mei-Ling
;
Hung, Yung-Hsiang
;
Lee, W. M.
;
Li, R. K.
;
Jiang, Bo-Ru
Recently, support vector machine (SVM) has excellent performance on classification and prediction and is widely used on disease diagnosis or medical assistance. However, SVM only functions well on two-group classification problems. This study combines feature selection and SVM recursive feature elimination (SVM-RFE) to investigate the classification accuracy of multiclass problems for Dermatology and Zoo databases. Dermatology dataset contains 33 feature variables, 1 class variable, and 366 testing instances; and the Zoo dataset contains 16 feature variables, 1 class variable, and 101 testing instances. The feature variables in the two datasets were sorted in descending order by explanatory power, and different feature sets were selected by SVM-RFE to explore classification accuracy. Meanwhile, Taguchi method was jointly combined with SVM classifier in order to optimize parameters C and gamma to increase classification accuracy for multiclass classification. The experimental results show that the classification accuracy can be more than 95% after SVM-RFE feature selection and Taguchi parameter optimization for Dermatology and Zoo databases.
Usage of Case-Based Reasoning, Neural Network and Adaptive Neuro-Fuzzy Inference System Classification Techniques in Breast Cancer Dataset Classification Diagnosis
(2012-04-01)
Huang, Mei-Ling
;
Hung, Yung-Hsiang
;
Lee, Wen-Ming
;
Li, R. K.
;
Wang, Tzu-Hao
Breast cancer is a common to females world-wide. Today, technological advancements in cancer treatment innovations have increased the survival rates. Many theoretical and experimental studies have shown that a multiple classifier system is an effective technique for reducing prediction errors. This study compared the particle swarm optimizer (PSO) based artificial neural network (ANN), the adaptive neuro-fuzzy inference system (ANFIS), and a case-based reasoning (CBR) classifier with a logistic regression model and decision tree model. It also applied three classification techniques to the Mammographic Mass Data Set, and measured its improvements in accuracy and classification errors. The experimental results showed that, the best CBR-based classification accuracy is 83.60%, and the classification accuracies of the PSO-based ANN classifier and ANFIS are 91.10% and 92.80%, respectively.