A Worker Recommendation Mechanism With High Acceptance Rates In Crowdsourcing Systems

With the thorough development of the Internet,mobile network and Internet of Things,crowdsourcing has emerged as a popular collaboration paradigm,in which tasks published by requesters can be economically and efficiently accomplished by crowd workers.Worker recruitment is an important aspect of crowdsourcing,worker recommendation has been widely adopted as an efficient way of worker recruitment.However,the existing worker recommendation only evaluates the trust that requesters place on workers.In the practical scenario,there are also untrustworthy requesters.Rational workers will refuse to participate in tasks issued by untrustworthy requesters,which will lead to the low acceptance rate of traditional worker recommendation mechanism.In order to solve this problem,based on the traditional worker recommendation mechanism,this paper further considers the existence of untrustworthy requesters,and designs a worker recommendation mechanism to evaluate mutual trust relationships between requesters and workers.The specific contents are as follows:(1)We propose a comprehensive and practical task matching mechanism by incorporating time matching,skill matching,payment matching,and location matching.To the best of our knowledge,no other previous work has considered in all aspects.(2)Based on biased matrix factorization,we propose three models to evaluate mutual trust relationship between requesters and workers.(3)By incorporating the task matching part and the mutual trust evaluation part,this paper proposes a Top-K worker recommendation mechanism based on greedy algorithm.(4)Extensive simulations and real data experiments highlight the performance of our proposed Top-K worker recommendation mechanism in terms of the trust evaluation and acceptance rate.