Research On Crowdsourcing Quality Control Strategies And Workers Reputation Model

With the fast development of Internet technology,crowdsourcing as a flexible and effective solution is receiving more and more attention.In recent years,great progress has been made in the field of crowdsourcing and many new applications that use crowdsourcing come into being.Different approaches are proposed,which has achieved good operating results.Since crowdsourcing is applied in complex online trading platform,the quality control of crowdsourcing application is a very important issue.The study of the way to effectively improve the quality of task completion,and the recognition of the quality of work completed by crowdsourcing workers has become urgent problems in crowdsourcing research.Meanwhile,the anonymous characteristics of crowdsourcing workers make its handling approaches different with traditional outsourcing tasks.Therefore,it is of importance to solve the quality issues of crowdsourcing accurately and efficiently.Crowdsourcing workers distributed widely,which attract map supplier to build up the geographic data collecting crowdsourcing based on smart mobile devices.However there are still many deficiencies in geographic information collecting crowdsourcing platform.For example,we cannot ensure the quality of the collected point of interest's information photos.This thesis makes deep research on the quality issues of the crowd sourced point of interest's information photos and proposed crowdsourcing audit policy for point of interest's information photos and the workers time reputation Model.Finally,the improved crowdsourcing auditing system was applied and tested in practical projects.The works finished in this thesis are as follows:1)Compare the different results in different crowdsourcing models,compare the existing crowdsourcing quality control policies and analyze their problems.2)Introduce the geographic information collecting crowdsourcing platform and propose crowdsourcing audit strategies for points of interest's information photos.Specifically,this work includes the design of crowdsourcing audit tasks,the selection of review workers and analysis of audit results.3)With the analysis of the shortage of choosing crowdsourcing review workers with Average credit,we proposed crowdsourcing workers time reputation model and analyze the value of parameters.4)Completed the points of interest's information photos crowdsourcing auditing system.5)Test the crowdsourcing auditing policy and workers time reputation model in audit work with different quantity of point of interest's information photos to verify the effectiveness of crowdsourcing audit policy and the workers time reputation model.Tests showed that the proposed crowdsourcing auditing policy for point of interest's information photo can effectively audit the quality of point of interest's information photo and organize better crowdsourcing workers.