Research Of Personalized Recommendation System Based On Object Feature Analysis

With the high speed development of information technology and the extensive using of the Internet,the big data era has already coming.Massive data on the Internet provide great convenience to widespread entrepreneurship and innovation,but at the same time,the problem of information overloaded becomes more and more obvious.It is more and more difficult for users to find their own data resources and information services.At present,collaborative filtering algorithm is the mainstream recommendation algorithm to solve the problem of information overloaded.However,the traditional recommendation algorithm is faced with the data matrix sparsityproblem and the problem of cold start.This paper proposes improvement plans from the following two aspect:Firstly,this paper proposes a method to calculate the similarity degree based on the multi-dimensional feature preference analysis.By intensive study,the origial single preference attribute information is divided into attributes in multiple dimensions.And then by using multi dimensional comparison method,obatian the comprehensive user preference information on multi angles,finally build the user association clustering.Secondly,this paper proposes a hybrid personalized recommendation algorithm based on features preference analysis in multi dimensional condition.First of all,obtain the effective information from the user's comments by Chinese Lexical Analysis System.Then recommend the similar feature relations clusters for target users by the method of multi dimensional user preference analysis.Summarize and analyze the characteristics of user preference and attribuion information.Fill the evaluation data matrix information by combining the traditional collaborative filtering recommendation model.Recommend the commdity object which is closest to the users preference attribution to the target users.This paper also design the a series of experiments to prove the effectiveness of the proposed similarity calculation method based on multi dimensional user preference.Compare with the traditional information recommendation algorithm,the personalized recommendation algorithm based on object feature analysis could effectively improve the accuracy of the recommendation informaion.The personalized information recommendation model based on object feature analysis proposed in this paper improves the shortcomings of the original algorithm.The sparsity of user data matrix and the newly registered user cold star problem are effectively alleviated.The new recommendation model provides system users with more accurate personalized recommendation service,users could easily find the data resources and information services.It could alleviate the problem of information overload in a certain extent,and has important practical value in the future application field...