Spam Host Detection Using Ant Colony Optimization

摘要

Inappropriate effort of web manipulation or spamming in order to boost up a web page into the first rank of a search result is an important problem, and affects the efficiency of a search engine. This article presents a spam host detection approach. We exploit both content and link features extracting from hosts to train a learning model based on ant colony optimization algorithm. Experiments on the WEBSPAM-UK2006 dataset show that the proposed method provides higher precision in detecting spam than the baseline C.45 and SVM.