Design And Implementation Of News Recommendation System Based On User Interest Drift And Semantic Features

With the rapid popularization and development of Internet technology,network resources are flowing in at an explosive speed.The development of the Internet makes it easier for people to obtain resources that they are interested in,but they also encounter the problem of information overload when enjoying massive amounts of Internet information.News recommendation system,as an effective method to solve the problem of information overload,can help recommend news that users may be interested in from massive news data.This paper researches the recommended methods in the field of news,mainly including the following work:(1)This paper proposes a news recommendation algorithm based on the user's dynamic interest perception.This algorithm calculates the weight of the influence of time factors on user preferences from the time series and content of the user browsing the news.Experimental analysis on real datasets proves that the algorithm has better performance than some baseline news recommendation algorithms.(2)This paper proposes a news recommendation algorithm that integrates multiple data information.The algorithm in this paper considers the impact on user preference from two aspects:temporal context and news characteristics.The algorithm proposed in this paper not only can effectively model the hidden factor model,but also has good scalability.Experiments on real datasets show that this algorithm can improve the accuracy of recommendation compared with other recommendation algorithms.(3)This paper proposes a news recommendation algorithm based on recency and relevance.This algorithm first mines the feature vector of the news,then models it based on historical reading behavior,and finally predicts the future-impact of the news.Based on this,Make non-personal recommendations.(4)Based on the algorithm proposed above,a news recommendation system based on user interest drift and semantic features is designed and implemented to effectively capture user interests and improve users' cold start problems,effectively improving user experience and recommendation accuracy.