Design And Implementation Of News Recommendation System Based On Spark

In today's world,information technology innovation is changing with each passing day.The wave of informatization is booming and global informatization is entering a new stage of development.In the age of information,the content of information is more abundant.At the same time,it brings negative issues such as “information overload”.A lot of information has brought certain problems to people's lives.In order to better solve the problem of “information overload”,the recommendation system becomes an important tool.In the process of R&D and application of the current news recommendation system,there are problems such as sparse rating matrix,unsatisfactory recommendation effect,and high real-time requirements.Aiming at the foregoing problems,this study constructs a news recommendation system that integrates multiple recommendation strategies and solutions to improve performance.Its feature is that new users can also obtain more accurate news recommendations;the news recommendations obtained by users meet their own forgetting curves;users can Get news recommendations quickly.This dissertation focuses on the recommendation algorithm as the core work.The paper solves the problem of data sparseness through the similarity transfer algorithm.The accuracy of user interest modeling is improved by integrating the time-weighted implicit semantic model.The real-time performance of the system is solved by the introduction of the Spark platform.Based on these theories can guide the construction and implementation of Spark based news recommendation system.The paper analyzes the functional and non-functional requirements of the news recommendation system from the perspective of software engineering,clarifies the design goals of the entire system,and designs the system's network architecture,function modules,and databases,and implements specific solutions for the system.Analyzed and analyzed the modeling process of news content,interest points,and personalized recommendations.The use of Spark platform and J2 EE platform for the implementation of the news recommendation system prototype deployment and testing.The news recommendation system researched in this dissertation has improved the real-time recommendation and sparse data.It has been verified that the experimental results are better.The research results can be applied to the actual recommendation service and can provide users with higher quality recommendations.Service will have strong practical value and practical significance.