Design And Implementation Of News Recommendation System Based On Hybrid Strategy

With the rapid development of the Internet and information technology,today's society has entered the era of information explosion.In the news,the rapid development of the Internet has made it easier to publish and read news on the Internet.Therefore,online news has become an important way for people to obtain information.Although the previous news websites have a large amount of news information,they only collect and merge the news,which leads to the users only passively receiving the news information provided by the news website and finding the required content,so the Internet has a large amount of and complicated.Online news information,but it is simply unable to meet the diverse and personalized news needs of users.In order to solve this problem,people are constantly looking for solutions,and the emergence of the recommendation system has become an effective measure to solve the above problems.Currently,the mainstream model of the recommendation system is the collaborative filtering model or the content recommendation model.However,there are two important problems in collaborative filtering.One is the cold start problem,and the other is the preference of the user of the item as the number of items and users increases.The matrix becomes sparse,and these two problems can seriously affect the recommendation accuracy of the recommendation system.In view of the above situation,the author has integrated this commonly used recommendation algorithm to construct a hybrid recommendation system.Through this system,not only can individual information be pushed for different users,but also to some extent,it can make up for the deficiency of a single algorithm.The newly constructed hybrid recommendation system pushes news of interest to the user to the user based on the user's demographic attributes,behavioral attributes,and hobbies.The content-based recommendation model utilizes the characteristics of news content,which is a textual content processing of news,recommending articles similar to the content for the user,and the model only recommends articles similar to the read news content to the user,but Did not take into account the user's own interests,the performance in recommending novelty is very poor.After the integration,we can learn from each other and greatly improve the performance recommended by the system.Through system testing,it is found that the hybrid recommendation system has different degrees of improvement in accuracy,recall rate,F value,per capita click,per capita exposure,next day retention rate and AUC compared with the recommendation system composed of a single recommendation algorithm.Therefore,the news recommendation system based on the hybrid recommendation mechanism has a very broad application prospect.