Web spam detection based on discriminative content and link features

The problem of spam detection is a crucial task in the web information retrieval systems. The dynamic nature of information resources as well as the continuous changes in the information demands of the users makes the task of web spam detection a challenging topic. So far many different methods from researchers with different backgrounds have been proposed to tackle with spam web pages problem. In this research, we study feature space of web spam detection to recognize most effective and discriminative features. Thereafter, we design a spam detection system that employs a minimum set of features and at the same time its performance is the same or very close to a system with the complete feature set. The experimental results show that we can reduce the number of features in a clever way while the accuracy of the system is intact or even improved.