Research And Implementation On Book Recommendation Algorithm Based On Collaborative Filtering And Latent Factor Model

Since entering the 21 st century,the Internet has been producing,creating and using data every day.Massive amounts of data generate a large amount of information,making it more and more difficult for users to distinguish useful information from it.The recommendation system can make personalized recommendations according to the user's interests,meet the specific needs of the user,and make the user and the product become more and more sticky.In the field of book recommendation,the recommendation system can make exclusive recommendations for different readers to meet the needs of fast and targeted search and increase the reading rate and sales of books.This dissertation focuses on the research and improvement of the recommendation algorithm based on collaborative filtering and Latent Factor Model,and applies it to the book recommendation field.Aiming at the collaborative filtering algorithm that does not consider the influence of the migration of user habits on the recommendation results,this dissertation integrates the time factor into the user-based collaborative filtering algorithm(User CF)and the item-based collaborative filtering algorithm(Item CF),constructs a user-based collaborative filtering algorithm(TF-User CF)and an item-based collaborative filtering algorithm(TF-Item CF)that incorporate time factors.Aiming at the Latent Factor Model algorithm that does not consider the attribute information of the user and the product itself,cannot handle the hidden data set and does not consider the user's cold start problem,the newly designed algorithm in this dissertation adds bias items and combines user attribute information,find the attribute neighboring users of the new user through the user attribute information,and then make recommendations.This dissertation uses the Movielens data set and Book-Crossing data set to conduct experiments to compare the average absolute error of the improved collaborative filtering algorithm and Latent Factor Model algorithm with the traditional algorithm.In the experiment of the improved collaborative filtering algorithm,it is found that the 7average absolute error will be reduced after the time factor is incorporated,which proves that the integration of the time factor into the collaborative filtering algorithm can improve the accuracy of the recommendation.In the experiment of the improved Latent Factor Model algorithm,it is found that the overall accuracy of both the traditional Latent Factor Model algorithm and the improved l Latent Factor Model algorithm is higher than that of the User CF algorithm and the Item CF algorithm.At the same time,the improved Latent Factor Model algorithm is more accurate than the traditional Latent Factor Model algorithm,which can improve the quality of recommendation.At the end of the dissertation,a prototype of a personalized book recommendation system is designed and implemented,which can make personalized book recommendations based on readers' behavior.