Crowdsourcing Quality Control Strategies And Valuation Algorithm Research

With the rapid development of Internet technology, crowdsourcing service as a flexibleand effective solution, more and more people began to be concerned about. In recent years,for all aspects of research crowdsourcing applications has made great progress, takecrowdsourcing means new ways of applications in many research fields, there are manyspecific approach, has been good treatment effect.Crowdsourcing applications generatecomplex background online network trading platform, there crowdsourcing application ofquality control problems, so study how to effectively improve the quality of task completionand malicious workers identified, to become the current crowdsourcing an urgent problem inthe research work. And the anonymous nature of the workers in the crowdsourcing platform,resulting in very different approach and traditional outsourcing tasks accurately andefficiently solve quality problems have important crowdsourcing.The paper first detailed introduce the causes and characteristics of crowdsourcingapplication generation, and compare it with the outsourced task differences. Thensummarizing crowdsourcing work mode and its applications in scientific research directionsand pointing out the reason for quality problems in the application of crowdsourcing.Furthermore introducing the advantages and disadvantages of existing quality control strategyand the processing method.Subsequently according to the current strategies and methods ofcareful research, found that the main research ideas to all of the crowdsourcing quality controlto take against malicious strategy. Therefore we attempt to change processing point of view,verifying crowdsourcing participants from the direction of the submitted data based on thecollection evaluation task.This paper is mainly about the relevance assessment task crowdsourcing results forquality analysis, put forward a kind of new crowdsourcing quality recognition model.Throughthe model of quality assessment method, to achieve the purpose of quality controlcrowdsourcing tasks, the experimental results show that the proposed model by theapplication of the method can better meet the need for crowdsourcing tasks presentedoutcome expectations, more accurate identification of workers type.