关键词: 竞争式众包, 设计众包, 任务模型, 能力评估, 推荐算法

Abstract: The massive personalized demand in the internet environment makes it harder to match design tasks and resources in crowdsourcing effectively.In order to improve the accuracy and efficiency of task matching in crowdsourcing contest,a two-way recommendation method considering the competency and participation motives of problem solver is proposed,through the analysis of the crowdsourcing design process.On the basis of content-based recommendation,this method firstly builds task model,problem solver's competency model and participation willingness model.Then the corresponding original set of tasks and original set of problem solvers are matched by task categories and skill tags of problem solvers.Thirdly the task from the original set of matched tasks will be recommended to the problem solver based on the participation willingness model.In the reverse way,the problem solvers from the original set of problem solvers will be ranked according to the evaluation results based on the competency model and the entropy method,and be recommended to the questioner.Finally,the experiment was carried out using the real data from Zhubajie.com,and the results were tested by the indicators of precision rate and recall rate.The results shows that the proposed method is better than label-based recommendation method and label-based recommendation algorithm for calculating match scores.This paper has a certain significance for solving the resource matching problem of massive personalized demand in crowdsourcing design.

Key words: crowdsourcing contest, design crowdsourcing, task model, capability assessment, recommendation algorithm