Personalized Comment Spam Filtering Using Clustering (In Korean)

Korean Institute of Intelligent Systems '08

Figure 1. System for comment spam filtering with clusters

As blog services are getting popular, spam comments are increasing abusing these. Comment filtering system basically relies on specific words or IP address/ID of already deleted comments by other bloggers. It could effectively prevent spams, however, suppress expressions and ignore different criteria for spams among bloggers. In this paper, we suggest a personalized spam comment filtering system which reflects bloggers' taste in comment.