The Research Of Credibility-Based Crowd Sourcing Tasks Allocation

With the development of the Internet,the crowd sourcing becomes an effective way of problem solving,the same time more and more people pay attention to it.The Crowd sourcing of the nature of the work to the workers within the platform are unknown group,using the collective wisdom to solve the problem,at the same time also brought a lot of uncertainty,this will make the quality control problem becomes one of the biggest challenges facing online crowdsourcing platform.The reason may be that the workers' interests are maximized,lack of professional knowledge,maliciously answered questions,and there is a bias to submit low-quality crowdsourcing results.Tasks at the same time,the traditional task allocation model based on the main demanders is conjecture and experience,because there are too many subjective factors,will lead to task allocation is not reasonable to reduce the utilization of human resources.In order to solve the problems,this paper proposes a model of crowdsourcing task allocation based on credibility,first of all,based on the historical record workers based on time window is introduced into active factor cut workers sequence,the credibility of the evidence obtained credibility;Then build credibility allocation model based on the above request and distribution model is the core of a weighted binary chart,in the heart of the binary chart released workers and demanders task as binary chart of the two kinds of nodes,respectively,using binary chart of maximum matching of weighted algorithm is obtained.Thus,the online dynamic matching of crowdsourcing tasks is realized.Each of the best matches is the best match between a task and a worker.Relevant studies have also proposed the crowdsourcing task assignment model,but there is no unified evaluation standard.This paper presents a general model for this problem,and tests the model with real experimental data and multiple evaluation factors.Compared with the existing credibility model,the validity and practicability of the crowdsourcing task distribution model based on credibility is presented.