Research On Information Recommender Systems Based On Information Evaluation Mechanism

Internet information is huge and growing rapidly,which brings a serious problem of information overload.Recommender system is an effective technology to deal with information overload,and has been widely used in business field.In this paper,we study the basic principle of recommendation system,and three popular algorithms:content-based recommendation and collaborative filtering,and the knowledge-based recommendation,their respective application scenarios and their pros and cons,as well as the evaluation indicators for recommended system performance.There has been relatively little proper RS research into demographic systems,but demographic-based recommendation is a good complement to basic recommendation algorithm,which is a starting point for this paper.The current recommendation algorithm relies on the characteristics of the user's historical interest,neglects the evaluation of the value of the information itself,thus triggering a series of phenomena of false information and inferior information.This paper proposes that the information evaluation mechanism can be introduced to improve the recommendation effect.On the basis of thorough investigation of previous theories and methods related to information evaluation,it is found that research is mainly focused on specific information sources or single information,and the value evaluation of information itself and new channel information such as WeChat official account and micro-blog are seldom studied.This paper proposes a mechanism to evaluate the information of different sources of the Internet,and completes three tasks:the definition of the evaluation object,the establishment of the evaluation index and the design of the evaluation method.Finally,based on changde city of hunan province,this paper designs a regional information recommendation system based on information evaluation mechanism.First of all,it is necessary to collect information to build the information base,and to establish static information library,semi-static information library and dynamic information base according to different information attributes.The information evaluation mechanism has three indexes.Relevant indicators refers to the regional classification of information,that is,the comprehensive use of Chinese word segmentation,place name recognition,rule matching and other techniques to complete the text information according to the city and county three level regional classification.The importance index is calculated by the weight of the source weight and the weight of the interaction with other sources and references,etc.Timeliness index is obtained through the relationship between the time attribute of information and the recommended running time.Finally,the regional information recommendation is realized.