Application And Research Of Deep Learning In Personalized Book Recommendation

Nowadays,with the rapid growth of information quantity,the problem of informat ion overload becomes more and more obvious.Recommendation system is an important method to solve this problem.It can help users find useful inform ation from a large number of data,so as to alleviate the problem of informat ion overload.Now,most online services such as e-commerce and social networking sites use recommendation engines to show their users the items that they are interested in.60% of You Tube's video playback comes from its home page recommendation,and80% of Netflix users' viewing comes from its recommendation system.In addit ion,the recommendation system has increased Amazon's profit by 35%.Therefore,the research in this field is very important.In recent years,due to the increase of computing resources such as GPU,the exponential growth of available data and the improvement of algorithms,deep learning has risen again,and has achieved good results in many fields such as computer vision and natural language processing.Firstly,this paper introduces the background and significance of the research in detail.Through the analysis of the background,the importance of using the recommended algorithm is obtained.The main work of this paper is put forward.Then it introduces the classification of the recommendation algorithm in detail,and gradually refines the research direction.It also introduces various applications of deep learning in the field of recommendation algorithm,such as: recommendation based on restricted Boltzmann machine,etc.In order to solve the problem that the accuracy of recommendation is not high because of sparse scoring data and relying solely on the traditional user item scoring matrix,this paper proposes a hybrid recommendation algorithm AI-DMF based on the matrix decomposition algorithm,which is composed of convolution neural network and depth neural network.The input model data introduces the auxiliary information(book name,book introduction,etc.)of the project,which makes the algorithm processing data more diverse,so that it can deeply mine the user characterist ics,help users accurately find their favorite books,and better recommend for users.The AI-DMF model is applied to two common data sets t o evaluate the model.When the data set data sparsity is very high,the AI-DMF can achieve a good score predict ion effect.