A study of online reputation system for blogosphere based on relationship and activity

Abstract

Reputation, as a part of collective systems of beliefs and opinions about people and things, is an important criterion to judge other people. With the rise of Web 2.0, reputation has gained a superior influence in large-scale online communities (Blog, SNS and etc) - which can be considered as a miniature of offline society- than it did in offline society. The online (digital) reputation encourages contribution and provides motivation for active participation, leading to better quality of the content and more user experience. Online Reputation systems are already being used in successful commercial online applications or recommendation applications. Nevertheless, people found because search engines repeatedly return currently popular pages or trendy keywords at the top of search results, reputable pages or blogs get ignored by an average user. Clearly, reputation is not equal to popularity or recommendation but people misuse these concepts. There is a rapidly growing literature around trust and reputation systems, but unfortunately this activity is not very coherent up to now. This research proposes model for blogosphere extracts online reputation that relies on relationship among people and activity of bloggers. The concept of reputation is start from humane studies or cultural science so reputation model in blogosphere is organized by existing methodology not in computer science but in sociopolitical field. Then, I make ranking by reputation model to verify the reputation model which I propose. A new ranking function is proposed to alleviate the problem of popularity based ranking. And the data included in metric can be gathered and analyzed through the web easily. For evaluating the proposed system, data are gathered through the blog in Korea consisting of a total of thousand pages. This research derives satisfaction and effectiveness through a careful analysis of a reasonable web data and experimental results that show the potential of reputation model...