Evaluation and Analysis of Popular Decision Tree Algorithms for Annoying Advertisement Websites Classification

Search engines are usually used for exploring the net and finding required information. When search results are shown usually 10 links are included in the first page. It must be notices how many percent of achieved results are related to our request. Unfortunately some of advertisement websites utilize phony techniques to attract users so that they could obtain their personal goals (such as increase in visit rate, higher rank, introducing products and so on). This type of websites are called annoying (intrusive) web pages which are sort of web spam. According to our study most of web users are not eager to see these pages. Moreover, these Web Pages waste users´ time and cause them to forget they search term as well as to fail in finding needed information. In this study various classification algorithms based on decision tree are evaluated and analyzed so that the best option for classification of these web pages is identified. The obtained results revealed that J48 is the best choice owing to its high precision and accuracy rate.