Research On Talent Recommendation Model Based On Project Tag And Its Application

With the substantial increase in the amount of service outsourcing in China and the promotion of the Internet + Traditional Enterprises boom,traditional advertising companies and practitioners about marketing services actively embrace the Internet,which promoted the rapid establishment and development of the Internet marketing platform in China.The efficient operation of Internet marketing platform relies heavily on high-quality talents and high-efficiency recommendation system;This study hope to improve the recommendation accuracy and efficiency about the talent recommendation system through project labeling.Thus reducing the enterprise's dependence on the experience of operators and improving the efficiency of the operation of the project management,thereby enhancing the competitiveness of enterprises.This paper firstly introduces the Concept and the basic principle of the recommendation system,and introduces the function and practical application of the label in the recommendation system,and then introduces the Content-based Recommendations Algorithm,User-Based Collaborative Flitering and Item-based Collaboration Filtering recommendation technology,and also summarized several algorithms in the research situation at home and abroad.With HYZX marketing platform as the research background,which introduce the basic situation,the project process and the present status of recommendation system about the HYZX platform.And points out the existing problems of the talent recommendation used by the platform.Though investigating the internet marketing platform,such as Pig Network,expert network and HYZX.Five key indicators(service category,the service object,industry,work style,project quotation)as Project label have been selected,calculate the weights of each label by relative comparison method,then constructs the basic model of talented person recommendation based on the Content-based Recommendations Algorithm and Item-based Collaboration Filteringrecommendation idea.This model is verified by a large number of random examples.Finally,combining the actual needs of HYZX,the paper designs the talent recommendation model to increase the receipt constraint.Finally,the basic model and the talent recommendation model of adding the single constraint is measured through the actual data of HYZX platform,and compared with the results of the proposed system,which verify the validity,Scientific and practicality about the recommendation model.