Fuzzy Support Vector Machine Based on Manifold Discriminant Analysis and Its Application on Web Text Classification
Support Vector Machine (SVM) is one of widely-used text classification method.Although SVM performs well in practice,SVM encounters two problems: the data distribution is not taken into consideration in the process of classification and its performance is greatly influenced by noises.In view of this, Fuzzy Support Vector Machine based on Manifold Discriminant Analysis (FSVMMDA) is proposed and Web text classification system is constructed based on Manifold Discriminant Analysis (MDA) and the fuzzy technology.The Web pages are firstly preprocessed including data cleaning, text segmentation and vector representation;And then,FSVM-MDA is used to classify on the prepared texts;Finally, the classification results are evaluated by the criteria such as precision,recall rate and the value of F1.The advantages of the proposed method are (1) it takes both the global and local characteristics into consideration;(2) it has the ability of noise-resistance.Comparative experiments on the authentic datasets verify the effectiveness of the proposed method.
web text classification manifold discriminant analysis fuzzy technology support vector machine
LIU Zhongbao ZHANG Jing
School of Software, North University of China, Taiyuan 030051, China
国际会议
武汉
英文
141-145
2016-10-21（万方平台首次上网日期，不代表论文的发表时间）