Fast web page categorization without the web page

Abstract

he World Wide Web has enormously increased day by day. Hence it is necessary for classifying the web pages. We demonstrate the usefulness of the uniform resource locator (URL) alone in performing web page classification. This approach is faster than typical web page classification, as the pages do not have to be fetched and analyzed Uniform Resource Locators (URLs) mark the address of the resource on the World Wide Web, are often humanreadable can indicate metadata about the resource[11]. Our approach segments the URL into meaningful tokens. We construct a binary tree for the entire set of tokens used in the hyperlinks and use J48 classifier. Our results show that in certain scenarios, URL-based methods approach show better performance.