Research On Recommendation System Based On The Label Semantics Framework

The rapid growth of the data and information promote the recommendation system technology research and development,the existing recommendation system can help people find their own information in the large-scale data to some extent,but the recommended system still can't accurately understand the needs of user.Simulation of human concept understanding algorithm is an important topic in the research and development of recommendation system.In a specified semantics context,the meaning of a vague concept is different from its literal meaning,which cannot meet the specified needs of the user.The focus of this paper is to mine valuable implicit information by existing text information in a specified context.In this paper,the application of fuzzy theory in the recommendation system is discussed,and the appropriate algorithm and the recommendation algorithm are analyzed.In the evaluation of appropriate degree,the label semantics framework is introduced,and it is a kind of subjective and open fuzzy semantics framework.In order to compute the appropriate degree of user,at the first,in a specified context,the participants pick out the appropriate label subset by the means of questionnaire,according to label subset picked out by all the participants,we can compute the appropriate degree of the corresponding logical expressions,which is corresponding to the appropriate degree.Recommendation system recommendation consists of three main elements: user,recommendation algorithm,and appropriate item.This paper utilizes registration information and the idea of label recommendation system to generate user portrait.According to the user portrait,we can recommend users with higher appropriate degree items.According to the above algorithm,in this paper,an experimental simulation based on the data of www.58.com,and a recommendation system based on label semantics framework is designed.First of all,the information of job seekers are analyzed,and the job seeker's information contains a job seeker's name,gender,age,work experience,current position,release time,we believe that the job seeker's age,work experience,educational background,current position can be used to evaluate a job seeker's ability.Secondly,through the crawl of www.58.com data,we find the current job attributes current position is not regular,in order to solve the problem,for not fill in the current position of job seekers to non-technical staff for processing,operator and technical operations to be classified as the technician category,and let the participants choose from processed data by the means of questionnaire.Finally,according to statistical result,the mass semantics value of different label subset is obtained,we utilize label semantics framework to compute the mass semantics value of user in different specified context and that is appropriate degree.Finally,in order to prove the effectiveness of the test,the test results are conducted from two aspects of accuracy and satisfaction.The experimental results and analysis show that the recommended system meets the requirements of user.