Research On The Recommendation Algorithm Of Online Book Store Based On Deep Learning

With the continuous development and innovation of various computer network information technologies,more and more data such as images,texts and other information resources in the Internet have been Shared more and more,which has also brought great convenience to People's Daily life no matter in the way of life,work or entertainment.In turn,the explosion of technology has brought about an explosion of information.People are immersed in a sea of information in a variety of different fields,and it becomes increasingly difficult to find the content they are really interested in or want to use through simple search and search.Therefore,in order to solve the problem of information overload,the recommendation system has become one of the most potential solutions.Currently recommended algorithm in the field of electricity is very wide,on major online bookstore buy books has become a widely used way of books,books most of them are in use on the Internet based on collaborative filtering recommendation algorithm,in the face of the cold start problems and data data sparseness slightly weak,online bookstore,the author of this paper try to improve the personalized recommendation algorithm,framework of book recommendation system,the improvement of algorithm are expounded,thesis studied mainly from the following several aspects:1.Establish the book corpus,modeled the natural language with the skip word model in Word2 Vec,obtained the word vector,and preprocessed the input data in the convolutional neural network.Compared with the existing corpus,the book corpus in this paper has improved its effect in preprocessing the input data of the book materials.2.The convolutional neural network model is improved on the basis of the existing model,and experiments show that the improved convolutional neural network model is slightly better than the original model in classifying books.3.Study based on the recommendations from the convolutional neural network technology and based on collaborative filtering recommendation algorithm in practice,effective combination,applied in recommender systems,recommend the combination of the improved results are superior to the traditional based on collaborative filtering recommendation algorithm,and to some extent alleviate the prevalent problems of cold start now traditional collaborative filtering algorithms.4.Implement a book recommendation system Web page,and adopted Flask Web as the front-end Web architecture,including the pages of login,registration,browsing and recommendation.My SQL is used as the back-end database to store customer and book information.