Research On The Reputation Evaluation Mechanism Of Contract Receivers In Crowdsourcing Based On Emotion Tendency

With the rapid increase of contract receivers in Crowdsourcing websites,how to reflect the ability of the receivers quickly and accurately has become an urgent problem to improve the Crowdsourcing process.Similar to the general e-commerce site,it is more common to reflect the reputation of the receivers through the comments of the users.However,there are differences between Crowdsourcing websites with e-commerce websites such as Amazon and Taobao,in Crowdsourcing websites,the number of comments is small,random and inaccurate,the existing reputation scoring mechanism cannot meet the needs of users.In this study,we define review text and comment scores together as user online reviews.Based on sentiment analysis of comment texts,comprehensive comment scores and other reputational influence factors such as comment time,task price,etc.,using Python web crawler technology,Based on the review data of zhubajie.com,the most well-known crowd-sourcing network platform in China,a set of appraisal mechanisms for the recipient's reputation in the crowd-sourcing model was constructed,and the rationality of the actual case verification mechanisms of the nine merchants was collected.The main research content is divided into:(1)Aiming at the fact that the crowdsourcing website already has a reputation system of the receiver,but the transaction experience of the contracted user is still not high,the connotation and characteristics of the crowdsourcing model and the signal theory are reviewed,and the Gibbons game model is used for analysis.The recipient's forgery capability information in the reputation system cheated on the task assignment behavior,and concluded that the reputation information should be used without fully understanding the ability of the recipient to provide a theory for the design of the recipient's reputation evaluation mechanism.basis.(2)Using the Python language to collect 95,364 user comments on the pig website,and use it as a corpus to build a sentiment dictionary after de-duplication and stop words.And in order to calculate the emotional sentiment score of the review text.(3)On the basis of dimensions such as total time,transaction amount and sentiment tendency scores,a set of reputation evaluation mechanisms for the recipients under the crowdsourcing model was aggregated,and the assessment was conducted in conjunction with the actual situation of the 9 outlets in the website of Zhubajie.com,to prove that this set of evaluation mechanisms can reflect the authenticity and ability of the recipient.