Research And Application Of Hybrid Data Clustering Recommendation Algorithm For Users And Scoring Information

In recent years,with the vigorous development of computer technology,the amount of internet knowledge has increased,and people have gradually come into contact with more information.They have gradually moved from an era of information shortage to an era of information overload.In order to quickly obtain the necessary information from the ocean of information,efficient information filtering methods are needed to process the information,and recommendation algorithms have emerged,and recommendation algorithms emerge as the times require.Recommendation algorithms are now ubiquitous across all platforms.When users browse and use the functions of various platforms,they leave their own historical behavior records for the platform.The content is provided to users,and personalized recommendations are provided for users.Currently,it is the most widely used recommendation algorithm that collaborative filtering recommendation algorithm,but there are still some problems,such as data sparsity,inaccurate similarity calculation,poor scoring effect,etc.These issues can have a significant impact on the quality of recommendation outcomes.This article studies the issues present in collaborative filtering recommendation algorithms.The main work includes:1.Aiming at the problems of inaccurate similarity calculation and poor scoring effect.In this paper,the user’s mixed data model is improved to calculate the similarity between users,so that the similarity can better reflect the similarity of users,so that the user’s prediction score can have more accurate results,thereby improving the quality of recommendation.This is because in the process of browsing and using various platforms,the user not only has a certain interaction process with the item,but also has a certain implicit interaction with the user’s own attributes,and the user’s attribute data is composed of symbolic data and numerical data.Composed,together with the user’s rating data,constitutes the user’s mixed data model,which requires a comprehensive analysis of the user’s mixed data model to generate a more accurate recommendation list for the user.2.Due to the significant difference in the number of users and items,the rating data tends to be very sparse,which results in poor performance of recommendation algorithms.This issue is commonly known as the problem of data sparsity.This paper improves this problem by performing cluster analysis on users,clustering similar users into one category,and then performing collaborative filtering recommendations on users in each cluster,using this method to divide user groups and mitigate the problem of data sparsity in recommendation algorithms,thereby improving the accuracy of recommendation algorithms.According to the similarity and clustering algorithm of the user’s hybrid data model,the hybrid data clustering recommendation algorithm(HDCRA)is developed to address these issues,and its effectiveness is verified through experiments.The experimental results show that compared with other recommendations,HDCRA has a certain improvement in alleviating data sparsity and improving accuracy.Finally,An improved algorithm is used as the basis for designing and implementing a library management recommendation system,and its practicality in application is verified through analysis of user data.