Performance analysis of artificial neural networks for automatic classification of web spam

Due to the increasing volume of information available on the web, search engines become increasingly necessary in day to day of Internet users. However, people with bad intention see this phenomenon as an opportunity to get profit and, consequently, a problem known as web spam is becoming increasingly common in the lives of Internet users, causing personal and economic losses. Fortunately, some methods have been proposed in the literature for automatic detection of this plague. However, the constant improvement of techniques used by spammers requires that the filtering approaches be more generic, efficient and with high capacity of adaptation. Well known techniques that have such characteristics are the artificial neural networks. Given this scenario, this paper presents a performance evaluation of multi-layer Perceptron artificial neural networks employed to solve such problem.