The Research Of News Topic's Subtraction Algorithm In Mobile Terminal Based On Contextaware Model

The recommendation system is one of the key directions in web applications.The user's preference and interest acquisition is one of the important research contents of the recommendation system.However,due to the disorder of the Internet,information overload occurred.When the user obtains news information,he/she will not be able to obtain the required information quickly,resulting in a bad reading experience.Therefore,the news recommendation system emerged at the historic moment,and continued to develop in the direction of precision and individuation.However,there are disadvantages that the recommendation condition is too single and the application of factors is small.For this reason,this paper decides to design and improve the recommendation system in terms of time context recommendation,improved the LDA algorithm,and algorithm fusion.And after designed is completed,the designed the recommendation system of the mobile phone is completed.Research includes:(1)Through improving the DTW algorithm,we design a ADTW algorithm with autocorrelation.The ADTW algorithm searches for repeated pattern patterns from related sequence data.Find out the law and find out the periodic pattern.The ADTW method can excavate the cycle without pre setting the cycle and solve the inaccurate mining cycle caused by noise items in time series.(2)Because the initial design of the LDA algorithm is designed for the English language,the use of the LDA algorithm in Chinese texts can result in ambiguous word storage due to inaccurate word segmentation.Therefore,in this paper,we introduce the method of improving the boundary entropy.(3)Most news recommendation algorithms nowadays use the time to analyze the news popularity in the application of the time context,thus affecting the degree of news recommendation.However,this will inevitably reduce the recommendation accuracy of users with periodic reading habits.Therefore,this article introduces the reference factor of user news reading cycle habits.Improve the recommendation accuracy in this way.(4)A set of news recommendation system based on time context is designed for the three improvements proposed in this paper.The overall framework of the design system,the design of each functional module and related classes,the entire forecasting system using PHP technology.After each improved design of the article,the proposed method was compared with the existing analysis methods.Through the test results,it is proved that the proposed method can really improve the accuracy of the recommendation.