Making the Most Cost-effective Decision in Online Paid Q&A Community

Yunhao Zheng,Xi Zhang,Yuting Xiao
DOI: https://doi.org/10.1145/3184558.3186349
2018-01-01
Abstract:Recommending proper experts to knowledge buyers is a significant problem in online paid Q&A community (OPQC). Existing approaches for online expert recommendation have been mainly focused on exploiting semantic similarities and social network influence, while personalizing recommendation according to individuals' motivations has not received much attention. In this paper, we propose a personalized expert recommender system, which integrates buyer's motivation for knowledge, social influence, and money in a unified framework. As an innovative application of cognitive computing, our recommender system is capable of providing users with the best matching experts so as to help them make the most cost-effective choice in OPQC. To this end, Paragraph Vector technique is implemented to construct domain knowledge base (KB) in a multilayer information retrieval (IR) framework. Then we perform knowledge pricing based on buyer's query and bid in the context of bilateral monopoly knowledge market. After that, a Markov Chain based method with user motivation learning is introduced to find the best matching experts. Finally, we evaluate the proposed approach using datasets collected from two OPQC. The experimental results show encouraging success as effectively offering reasonable personalization options. As an innovative approach to solve the expert matching problem in OPQC, this research provides flexibility in customizing the recommendation heuristics based on user motivation, and demonstrate its contribution to a higher rate of optimal knowledge seller-buyer matching.
What problem does this paper attempt to address?