| This paper mainly discusses the book recommendation technology based on collaborative filtering algorithm,and combines the clustering algorithm and user characteristics on the basis of collaborative filtering,proposes a collaborative filtering recommendation algorithm based on user clustering and user characteristics,and finally designs and implements a book recommendation system based on clustering and user characteristics.The main work includes the following four aspects:(1)Research the current main recommendation algorithms,among which collaborative filtering recommendation algorithm is the most common and typical personalized recommendation algorithm.The basic idea is to mine the user’s historical behavior through data,and predict the user’s upcoming behavior based on the data to generate recommendations.In e-commerce platform,collaborative filtering algorithm is the most widely used algorithm.Traditional collaborative filtering algorithm still has many problems that can not be ignored,such as data sparsity,cold start,etc.In order to solve these problems,this paper proposes to improve the traditional algorithm.(2)A collaborative filtering algorithm based on clustering algorithm is proposed to improve the data sparsity problem by improving the user similarity.Although the algorithm has solved the problem,the calculation of the similarity between users is large,which has a great impact on the algorithm in general.The improvement of cold start is not obvious,and the algorithm needs to be improved.(3)In the collaborative filtering algorithm based on improved clustering algorithm,user characteristics are also a particularly important factor when generating the nearest neighbor set.Different users have different degrees of interest in books,indicating that user characteristics may be the same.In the aspect of algorithm improvement,users can be classified first through user characteristics,which will greatly improve the accuracy of the nearest neighbor recommendation.(4)A book recommendation system based on clustering and user characteristics is designed and implemented.On the basis of the improved recommendation algorithm,apply it to the book recommendation system,design the overall framework and main modules of the system,and design a recommendation system with beautiful interface,simple use,practicality and consistency,which can realize book recommendation to users. |
