Comparison of Neural Networks and Support Vector Machines Using PCA and ICA for Feature Reduction
Web page classification provides an efficient information search to internet users. However, presently most of the web directories are still being classified manually or semi-automatically. This paper analyses the concept of the statistical analysis methods known as Principal Component Analysis (PCA) and Independent Component Analysis (ICA). The main purpose for using integration of PCA and ICA in Web News Classification is to perform feature separation and reduction. The feature vectors are applied to Neural Networks (NN) and Support Vector Machines (SVM) classifiers. F-measure is used to measure the classification effectiveness and found SVM is better than Neural Networks (NN). For the classification-ability experiment, sports news web page section was used.