Design And Implementation Of Automotive Article Recommendation System Based On Hybrid Recommendation Mode

With the rapid development of the automotive industry and the improvement of people's economic level,automobiles have increasingly entered ordinary families.The demand of people for automotive articles is also increasing day by day,while with the rapid development of the mobile Internet,the automotive media also shows a great vitality.Based on the above situation,the development of automotive article recommendation systems based on automotive information-based mobile phone application will not only helps users extract content of interest from the mass car information,but also brings benefits to the automotive information-based mobile phone application.Therefore,a win-win situation for users and enterprises has been achieved.With the continuous development of the recommendation system,a variety of recommendation strategies emerged.Each recommendation strategy has its own advantages and disadvantages.Therefore,in order to make up for the inadequacies of various recommendation strategies,in this project,we use hybrid recommendation mode to combine multiple recommendation strategies to produce recommendations for users.The specific recommendation strategies include content-based recommendation,collaborative filtering recommendation,and association rules-based recommendation.In particular,the content-based recommendation is divided into keyword-based feature recommendation and vehicle tag feature recommendation based on the different content characteristics.The author completed the requirement,analysis,design,implementation and test of the data synchronization and processing module,the recommendation strategies off-line computing module,the recommendation engine module and the recommendation background management module of the automotive article recommendation system based on the hybrid recommendation mode.In particular,the core module of the recommendation system is the recommendation strategies off-line computing module,which uses the MapReduce computing framework to implement the distributed parallel computing for a variety of recommended strategies.The recommendation engine module uses the Spring Boot framework as the server framework,and then implements personalized recommendation and related recommendation functions combined with the results of recommendation strategy off-line computing.The data synchronization and processing module realizes synchronization and processing of multiple data,then provides data foundation for user interest mining and calculation of recommendation strategies.The recommendation background management module takes the recommendation effect monitoring as the core function,then uses ECharts to realize the visualization of the recommendation effect data and the management of the relevant data of the recommendation system,so that the developer can improve the recommendation system according to the recommendation effect feedback.In this project,the hybrid recommendation model based automotive article recommendation system is combined with the large data platform to establish an exclusive interest model for users through the analysis of user historical behavior,so as to provide exclusive recommendation services for users in the era of serious information overload,and this project is committed to promoting user satisfaction and click through rate of automotive information-based mobile phone application.