Article ID Journal Published Year Pages File Type
10312568 Computers in Human Behavior 2015 11 Pages PDF
Abstract
Online forums have been extensively used in many organizational knowledge management practices as well as virtual communities for sharing knowledge and opinions. Identifying experts in certain domains is essential for improving knowledge sharing and accessibility through online forums. Existing expert identification techniques can broadly be classified into two major approaches: content-based and link-based. Although the link-based approach has shown its superiority over the content-based approach, it incurs some limitations when applying to the task of identifying experts in online forums. In this study, we propose an expert identification technique that relies on the opinion ratings from the members in an online forum. Specifically, we extend PageRank and propose the ExpRank algorithm, which considers both positive and negative agreement relations among the members of the online forum. Using two datasets (pertaining to different product categories, books and music) collected from a well-known product-review website (i.e., Epinions.com), our empirical evaluation results show that our proposed ExpRank algorithm outperforms its benchmark technique (i.e., PageRank). Our evaluation results also highlight that the incorporation of negative agreement relations can improve the effectiveness of expert identification.
Related Topics
Physical Sciences and Engineering Computer Science Computer Science Applications
Authors
, , , , ,