Personalized Recommendation Method Based On Multidimensional Scene Feature Classification

Nowadays,Internet intercommunication involves all walks of life,such as the medical,education,aviation,industry and so on.The amount of data that can be crawled on the network,the amount of data that needs to be stored and analyzed is increasing exponentially,and the shared data are more and more.Therefore,the problem of information overload is becoming ever more serious.Therefore,the recommendation system is in the big collar.The domain is becoming ever more popular,and the demand for recommendation methods in industrial applications is higher and higher.Most of the recommendations are built on the interactive information of users and products and rarely join other related information.The recommendation method based on multi-dimensional scenario information will consider the impact of multidimensional scenario information on the future behavior of users in the recommended process.The personalized recommendation method in this paper dynamically extracts multi-dimensional situation features of users and products,analyzes the user's interest in products,and the hidden relationships between users and products,products and products,users and users,thus greatly improving the user's satisfaction with the recommended results and the effectiveness of accurate marketing.Based on this,this paper proposes a personalized recommendation algorithm based on multidimensional scene feature classification.(1)Reading a large number of relevant literature,summarizing the achievements and shortcomings of the previous research,the theoretical basis and fundamental technologies needed to be used in this paper are found and summarized in a large number of data.This paper proposes a multi-dimensional scenario classification for users and products,which can provide users with more accurate and effective product recommendation list and provide more accurate marketing objects for the product,that is,the user list,and can analyze the customs relations between users in the same situation,and the relationship between products and products.Joint relationship.(2)Clustering the multidimensional scene features of users and products,and building a classification model.A scenario diagram of user product user relationship is constructed based on collaborative filtering.On this basis,a user interest degree algorithm for multi-dimensional situation classification is constructed,and a personalized recommendation method for users and a recommended method for product precision marketing is designed.(3)The simulation experiment is conducted.The scene features of the user and the product is extracted from the experimental data.The Single-Pass clustering idea is used to cluster the user and the product.Then the decision tree and the random forest are utilized to train the classifier for the user and the product category label after the clustering.The personalized recommendation method is classified,and the experimental results are evaluated,compared and analyzed to test the effectiveness of the proposed algorithm.