Recommendation Method Of Product Design Task For Crowdsourcing Contest

With the globalization of economy and the development of information technology,the crowdsourcing model of product design based on internet platform has been developed rapidly.Effectively improving the matching efficiency between design resources and design tasks has become an urgent problem to be solved.At present,most of the crowdsourcing contest task recommendation algorithms focus on recommending tasks to problem solvers according to their interests and hobbies,while few recommend problem solvers to tasks.Moreover,the traditional content-based recommendation has problems such as difficult to distinguish the quality of recommended content.Collaborative filtering has the problems of high online running time and no copy of crowdsourcing tasks.To solve the above problems,this paper proposes a two-way recommendation method that not only considers the personalized needs of problem solvers and problem proponents,but also can distinguish the ability of problem solvers and reduce the online running time of algorithms.Firstly,the crowdsourcing contest model is studied,a process suitable for crowdsourcing contest product design is proposed,and the personalized needs of problem solvers and problem proponents are analyzed.Secondly,based on the modeling process of content recommendation method,a recommendation method suitable for the crowdsourcing process is established.The recommendation method includes PS-to-QR and QR-to-PS,as well as crowdsourcing task model,problem solver ability model and problem solver participation motivation model.Secondly,formulate the recommendation rules.The recommendation rules of PS-to-QR are to first obtain the PS set that can complete the crowdsourcing task based on the skill tag and task category matching,and then introduce the entropy weight method to sort the capabilities of the PS set to form the Top-n recommendation list;The recommendation rule of QR-to-PS is to first match the task set that meets the PS participation motivation based on the skill tag and task category,and then form the Top-n recommendation list according to the PS participation motivation priority.Finally,according to the real-time data collected by Z crowdsourcing platform,the experimental verification is carried out,and the experimental results are compared and analyzed by using indicators such as precision,recall and mean average precision.The experimental results show that the designed recommendation method is better than the tag based recommendation method and the tag based recommendation algorithm for calculating the matching score,which can effectively improve the resource matching efficiency of the personalized demand of sea volume in crowdsourcing design.Based on the theory and method research of this paper,a recommendation system integrating crawler and recommendation is designed and developed.The real data is obtained from Z platform,which can realize the two-way matching recommendation of design resources and task resources.