Text Category Crowdsourcing Solution Filter Research In Reward Mode

The crowdsourcing is a new network model under open innovation environment.With the rapid development of the Internet,crowdsourcing model has been favored by enterprises,emerging many crowdsourcing network platform,such as the AMT,Zhubajie,TaskChina and so on.The reward mode is the main mode of crowdsourcing,in which a crowdsourcing task can collect a number of crowdsourcing solutions.However,the task promulgator just chooses the appropriate solutions.Nowadays,the choice of crowdsourcing solutions is often carried out in manual way,but low efficiency,high cost and strong subjective.How to choose suitable solutions from a large number of crowdsourcing projects is a problem that needs to be solved urgently.This paper studies the selection of text category crowdsourcing solution in reward model from the perspective of text information filtering,and constructs a double layer filtering model of crowdsourcing solutions.The purpose is to realize the initial screening of the solutions in crowdsourcing network platform.The filtering model consists of three functional modules: the crowdsourcing text data preprocessing,the double-layer filter based on the Rocchio algorithm and the filtering processing module.The preprocessing of crowdsourcing data is carried out in four steps: firstly,the task and solutions crowdsourcing text data are collected.Secondly,the Chinese word segmentation tool NLPIR is used to segment the text,and a series of feature words with frequency statistics are obtained.Thirdly,the feature words are extracted according to the word frequency of the document,and the feature matrix is established.Finally,the vector space model is used to describe the contents of the crowdsourcing text.In order to verify the feasibility of the filtering model,it is used to filter the actual crowdsourcing solutions on the crowdsourcing network platform,and the results are compared with the artificial selection.It shows that the proposed filtering scheme has a certain validity and applicability,which can provide a feasible and theoretical basis for realizing the crowdsourcing solutions automate initial screening and it has certain practical significance and realistic value.