In recent years,crowdsourcing has developed rapidly as a new business model and has been widely used in many fields.However,crowdsourcing faces an uncertain organization of the mass population,workers in the crowdsourcing markets usually have different background,expertise and incentives.The anonymity of the Internet makes the information dissymmetrical between the two parties,therefore,workers are likely to exhibit heterogeneous quality in their submitted work.Therefore,how to control the quality has become a hot issue in the field of crowdsourcing.The article summarizes a large number of domestic and foreign research literature,introduces the concept of crowdsourcing model and the related research work about crowdsourcing quality control,and analyzes the advantages and disadvantages of the quality control strategy of crowdsourcing.Through the actual investigation of crowdsourcing platform,we find that there are a lot of workers' historical trading behavior on the platform.It is possible to predict the future behavior of the workers through the historical information and reduce the information asymmetry between the two parties.The article proposes a quality control method for crowdsourcing based on workers' reputation,mainly including crowdsourcing quality assessment and dynamic incentive.Firstly,evaluate the quality of crowdsourcing.A quality evaluation method is proposed to consider the reputation of workers.First,build the reputation model based on the workers' past performance.Then,select the workers according to the reputation of workers and put the reputation as the weight into the EM algorithm optimizing the initial value selection method,preparing for the quality control of crowdsourcing.Then,based on the evaluation results,a dynamic incentive method based on the quality of the workers is proposed.This method is a two-stage dynamic excitation.Set up the task segmentation point and implement the unified payment mechanism in the first stage,attracting enough workers to participate in the task.Set the dynamic excitation function in the second stage,which is positively correlated with the quality assessment results of the workers and is negatively correlated with the task completion.Finally,verify the validity of this method through numerical experiment and rate the threshold of parameters.The experimental results show that the quality control method proposed in this paper can improve the quality assessment results and motivate workers to submit high quality results. |