The Design And Implementation Of Personalized News Recommendation System

The News of international,social,people's livelihood and other different kinds produced every day are complex and diverse,and the news based on the week is even more numerous.The information overload news makes people overwhelmed.How to extract the news that users really care about has become an urgent need at present.Under this background,news recommendation system can recommend topics and news that may be of interest to users by extracting user interest features,which has practical significance.In this paper,the main work of personalized news recommendation system is as follows:Firstly,according to the input needs of news recommendation algorithm,we classify and labels news,and realize news preprocessing,classification,labeling and news recommendation system as a whole and investigate the modeling of news text processing and classification.News exists in the form of hypermedia containing pictures and videos and can't be directly used in the calculation of recommendation algorithm,by extracts news text,processes text keywords,and confirms the final news classification algorithm.We determine that CNN algorithm is to be used as the news classification algorithm in this paper from the current classification clustering algorithm.Secondly,by studies several current recommendation algorithms and confirms last recommendation algorithms.According to the characteristics of news recommendation,users can not only pay attention to the same kind of topics they like,but also may be interested in the current hot topics.Therefore,taking heat-based recommendation algorithm into account,a hybrid recommendation algorithm based on label,heat and collaborative filtering is designed for news recommendation.Current recommendation algorithms mainly include content-based recommendation algorithm,collaborative filtering recommendation algorithm,hybrid recommendation algorithm and label-based recommendation algorithm.This paper studies and combs various recommendation algorithms,confirms that this paper intends to use hybrid recommendation algorithm as personalized news recommendation algorithm in this paper.Finally,we design,implement and test the personalized news recommendation system,design the overall framework of the recommendation system,implement the import and classification of news content,implement the recommendation algorithm using the interface provided by Python,and test the system integration,including the news text preprocessing,the linear weighting of news text phrases,the correctness and usability of the text classification algorithm,and tests the validity and usability of the algorithm.The hit rate and efficiency of the recommended algorithm.Experiments show that the news recommendation algorithm designed in this paper has high hit rate and great practical application value.