Research On Personalized Recommendation System Based On Tag

With the rapid development of computer and Web2.0 technology,there's a huge amount of information on the Internet.In the face of so much good or bad information,it is difficult for the users to extract their own real need for information resources effectively,which is what we call "Information Overload" issue.In order to solve this problem,people have created categories and information filtering tools,such as searching engine,web portals and so on.However,these tools are not enough to meet the needs of individual users,because of the need,personalized recommendation system was born.In recent years,personalized recommendation algorithm has become the focus of academic research.However,the traditional recommendation algorithm has some limitations.Besides,it is difficult to achieve personalized recommendation with high speed and accuracy of specific applications.To solve this problem,a recommendation algorithm based on hierarchical tag is proposed,which is called HTRA(Hierarchical Tag Recommendation Algorithm).The main work of this thesis is as follows:First of all,when adding the tag for items,this thesis uses the del.icio.us data set to demonstrate the cosine similarity calculation,which can get higher accuracy when calculating the tag similarity.Secondly,the shortcomings of the traditional tag recommendation algorithm are deeply analyzed,and a hierarchical tag recommendation algorithm is proposed.The algorithm uses the tag data to characterize the feature information of users and items,so as to get the user preferences of historical items.Finally,the similarity between the historical items and the items to be recommended is calculated,and the adjustment factor is introduced to combine the user's interest preferences and the object similarity.The user's preference value of the product is predicted.Simulation results show that the proposed algorithm is effective and the accuracy of the proposed algorithm is improved.According to actual requirements for online advertising recommendation projects in association enterprises,this thesis designed and built the overall framework of personalized recommendation system,and designed the corresponding functional modules in detail through the detailed analysis of the functional requirements and performance requirements of the recommendation system.By using the My Sql database,an advertisement recommendation system is implemented based on Node Js.And the system implementation mainly includes three parts: anchor data preprocessing,the HTRA algorithm implementation and the advertisement recommendation system implementation.The anchor data preprocessing is based on the rules defined by the system to extract the more popular anchor information as the system user.The algorithm uses the HTRA algorithm in this thesis and the website implements the main functions of the system,including registration,login,advertisement and recommendation function.Finally,the function and performance of the system are tested and analyzed.The results show that the proposed system has good recommendation quality.