CredibleExpertRank: Leveraging Social Network Analysis and Opinion Mining to Facilitate Reliable Information Retrieval on Knowledge-Sharing Sites

Abstract

Asking and answering questions are common activities in both the workplace and everyday life. Knowledge-sharing websites have become a popular resource for obtaining instant and searchable answers. However, users of these sites may encounter challenges in acquiring timely and appropriate content from user-provided answers owing to factors such as limited expertise, spam, and time constraints. Identifying trustworthy experts who can provide relevant and reliable answers in knowledge-sharing communities is crucial to overcome this issue. In this study, we propose a solution to the problem of identifying credible experts on knowledge-sharing sites by introducing the CredibleExpertRank algorithm. Our algorithm calculates a CredibleExpert score based on two main factors: activity and credibility. The credibility score is determined by analyzing users' interactions related to questioning, answering, recommending, and mining users' opinions, while the activity score reflects the user's level of participation on the platform. We conducted experiments to evaluate the performance of the CredibleExpertRank algorithm, using user satisfaction measures for answers to given questions. Our findings confirmed that the credible experts identified by our algorithm provided more relevant and timely answers compared to other ordinary users. The timely nature of the credible experts' answers was due to the reflection of their activity factor, while the superior performance in relevance was attributed to the high recommendation rate of their answers and positive evaluations received from opinion mining results. Our study undertakes an extensive investigation focused on the identification and prioritization of credible experts, revealing their profound advantages in significantly enhancing the overall quality of knowledge-sharing platforms. We proposed the CredibleEx-pertRank algorithm as a powerful method for effectively identifying trustworthy experts and giving priority to their answers. Through a meticulous process of experimental evaluation, we provide compelling evidence that this approach leads to substantial improvements in both search efficiency and reliability on knowledgesharing sites. By highlighting the potential benefits derived from the identification of credible experts, our study underscores their pivotal role in elevating the overall performance of knowledge-sharing platforms.