Performance Evaluation of a Countering Method to Social Bookmarking Pollution Based on Degree of Bookmark Similarity

Abstract

Social Book marking (SBM) is one of the most widely used Web services. An SBM website displays and shares each user's bookmarks. The SBM service aggregates the number of users who bookmark a given Web page and provides useful information as a result of these aggregations. However, an increase in the popularity of the SBM service and in the number of the users of the SBM service results in an increase in the amount of SBM SPAM. In addition, the SBM service generates irrelevant information to many users because of the aggregation of a large number of bookmarks, we call this problem "SBM pollution." In this paper, we propose a method for countering the problem of SBM pollution based on the degree of bookmark similarity. The proposed method creates blacklists that contain lists of users having a high degree of bookmark similarity. Based on the created blacklists, the number of bookmarks of the Web pages influenced by SBM pollution is reduced. From the results of the performance evaluation, we show that our method reduces the number of bookmarks of most Web pages influenced by the SBM pollution to a great extent.