期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2014
卷号：5
期号：6
页码：7339-7345
出版社：TechScience Publications
摘要：In the online social network such as Facebook, Twitter, etc., it is possible to display any type of data on the wall of the user. These data can contain unwanted messages such as: policy statement, vulgar data, account staff teasing, etc. other users can see the data and may also rule on such a position. Such a position can affect user social image. So, the safety of such personal wall is an important issue. To some extent face book allows users to specify which helped to put messages in their walls (i.e., friends, friends of friends, or defined groups of friends). However, no preferences based on content are supported, so it is impossible to prevent the display of these unwanted messages. To protect a spam message posted on the wall of the user and to protect the user social image is an important issue in the social networking site. To filter out unwanted messages, we offer three levels architecture containing a message classifier based on the content and use of machine learning technique. The user is able to customize the filtering rule according to their preferences and / also set the filter on different user privileges i.e. subsidies to allow the user to insert messages in his / her wall. We propose a system that allows the user to restrict particular user based on his social reputation also user can extract tags from images displayed and filter accordingly and make the decision whether to allow such content or not.
关键词：online social network; learning machine; short text;classification; reputation.
