Webspam Detection Using Classification Algorithms and Optimizing the Performance of Classifiers by Selecting the Features

Web spam is one among the main problems of search engines because it reduces the standard of the online page. Web spam also effects economically because spammers / attackers provide an oversized free ad data or websites on the search engines that results in a rise within the web traffic. There are certain ways to tell apart such spam pages and one among them is using classification techniques. Relative examination of web spam detection using classification algorithms like, Random Forest and LAD Tree, C4.5 and Naive bayes is presented here during this paper. Analyses were completed on highlight sets of all around acknowledged dataset WEB SPAM UK-2007 utilizing WEKA. When classification was refrained from feature selection some classifier were high on false positive rate and time taken to create model but when feature selection was applied to datasets results were optimized and Random Forest outperformed on all the datasets altogether parameters that were selected.