Online News Text Classification using Neural Network and SVM

Abstract

Categories for classification of text are predefined according to these categories all text data is classified. We require classifying text to manage and search any data in database. There are many techniques available in market to classifying the text. Now days every website has overloaded text in database as like customer support websites, news website etc. so in this type of websites text need to classify. In news websites it’s necessary to maintain record of old and new news into the database. The news can be classifying on the basis of predefined categories of type crime news, sports news, election news, entertainments etc. every technique that exists in real like SVM, Naive Bayes, and Neural classifiers, working well at a level with some limitations. In this we are going to discuss about these techniques and conclude with the comparison of results find out which technique can perform well. Text category detection refers to identifying the type of category getting used by the text. The process involves two process training and testing. The training section involves the feature extraction process and the testing section involves the identification of the type of text used. In the process classification, involvement of a classifier is there to check the accuracy of the training. In this we focuses on the enhancement of the text category detection using back propagation neural network and Support vector Machine. The classification results have improved by 5 to 10 percent