A meaningful model for computing users' importance scores in Q&A systems

This paper presents a meaningful model for computing users' importance scores in Q&A systems that can stimulate the development of these systems. Since the score can be used in ranking users' expertise, it can motivate users to contribute more to Q&A systems. Direct methods to rank users such as counting number of answers, number of votes and so on do not work because they provide limited incentives, and spammers can make use of them quite easily. Our model, which is based on random walker model and referred to PageRank algorithm, indirectly calculates the importance score of each user through his/her relationships to other ones of Q&A systems. This method does not only limit spamming hazards but also provide users with some additional incentives to contribute more to the systems. Last but not least, the algorithm is designed to be able to be implemented on MapReduce programming model so that it can be applied to large-scale systems.