Personalized Recommendation Technology Research And Implementation In RSS Network News

As the time for data concentration is coming,how to find what we need from overload information quickly and accurately has become a hot topic.The appearance of RSS technology changes the way in which people obtaining information.However,the research of RSS technology focuses on how to extract the RSS feed,ignores the requirements of individual,and reduces the quality of service.Most of existing individual research focuses on recommendation for books,electronic commerce,audio,and food.But a few research achievements on personalized recommendation for news.In the news item recommendation systems,usually use TF-IDF weighting technology combined with the cosine similarity measure,however,this technique does not take into account the actual semantics of the text itself.Therefore,after analysis the traditional personalized recommendation technology and compare to the news characteristics,this topic proposes are commendation method based on semantic analysis of news content.Firstly,review the current personalized recommendation algorithm and RSS technology development status,and comparative analysis and compare the mainstream recommendation algorithm.Secondly,the current content-based personalized recommendation algorithm is mostly based on vector space model.Not only the prevalence of high-dimensional sparse,but the more important question is the nature of news is text data.Text data has the specific feature that natural language problem,which has not been well utilized in the current recommended method.Therefore,this paper uses a new approach,via introducing of semantic analysis;the news text content is expressed as a number of synonym synsets.And improve user interest model,using semantic similarity calculation the similarity of resource information and user interest model.Then,In order to get more accurate results,after study of several classic semantic similarity calculation methods,we propose a method based on extended semantic relationship tree,using this method to calculate the similarity of resource information and user interest model,Verify algorithm performance has improved through experiments.Finally,use the proposed recommendation algorithm based on semantic analysis recommends RSS news,Combine the personalized recommendation technology and RSS technology.Verify the improved model and the performance of algorithm by simulation experiments,compare with traditional content-based recommendation algorithm.Experimental results show that recommendations based on semantic analysis of news is feasible,and enhances the accuracy and efficiency of the recommendation.