A Generic Approach for Web Page Classification Using URL’s Features Along With the Textual Content

Classification of web pages greatly helps in making the search engines more efficient by providing the relevant results to the user’s queries. In most of the prevailing algorithms available in literature, the classification/ categorization solely depends on the features extracted from the text content of the web pages. But as the most of the web pages nowadays are predominately filled with images and contain less text information which may even be false and erroneous, classifying those web pages with the information present alone in those web pages often leads to mis-classification. To solve this problem, in this paper an algorithm has been proposed for automatically categorizing the web pages with the less text content based on the features extracted from both the URLs present in web page along with its own web page text content. Experiments on the bench marking data set “WebKB” using K-NN, SVM and Naive Bayes machine learning algorithms shows the effectiveness of the proposed approach achieving higher accuracy in predicting the category of the testing web pages. Our proposed algorithm achieves higher accuracy 90% when K-NN is employed on the given data set. There is also considerable improvement in accuracy using other two algorithms when employed for classifying the web pages based on our proposed approach.