| With the advent of the new era of the Internet,computer technology and application ecology have entered a stage of high-quality development.When users enjoy the big data dividends brought by the network,they also face problems such as information overload,which makes it impossible for users to quickly and clearly screen out valid data or information of interest.The recommendation system described in this paper can actively recommend personalized information for users,thereby improving user satisfaction.Traditional recommender systems face problems such as data sparsity and cold start,which restrict the recommendation performance of the system.The purpose of this paper is to build an efficient recommendation model that integrates knowledge graphs and apply it to the music field,that is,to design and implement a personalized music recommendation system.Firstly,based on the research background of the subject,this paper introduces the development and research status of recommendation methods and knowledge graphs,expounds common personalized recommendation methods,and summarizes the main construction techniques of knowledge graphs.By learning these relevant theoretical knowledge and techniques,it lays a solid foundation for the follow-up research work.This paper adopts the constructed domain knowledge graph to improve the quality of the knowledge graph.The main content is to use the selected ontology description language to build the knowledge graph ontology,then use the script to obtain data and store it in the database,and then perform knowledge mapping and serialization on the database according to the constraints of the ontology.Finally,the serialized knowledge is stored in the graph database to form a high-quality domain knowledge graph.Then this paper designs and implements a personalized music recommendation system.The main content is to analyze the requirements of the system,and then design the overall architecture,subdivision functions and database tables of the system.Then according to the overall design,we developed and programmed to realize the functions of each module.Finally,deploy the system and conduct functional and performance tests.The core research of this paper is to build a personalized recommendation model that integrates knowledge graphs,including model-building recommendation tasks,knowledge graph tasks,and building feature fusion modules.By training the model,designing various comparative experiments,etc.,the experimental data is analyzed graphically according to the evaluation criteria.The research results show that the model in this paper has a certain degree of performance improvement in click-through rate and Top-N recommendation.Finally,this paper applies the constructed model to the developed music system to complete the music recommendation. |
