Study On Intelligent Recommendation Technology And Application In E-Commerce

The popularization of E-commerce applications brings in not only the rapid development of E-commerce but also the overload of information. As an important tool of E-commerce, the recommendation system enables users find useful purchasing advices in mass merchandise information to improve Sales, which ensures the recommendation system a fine application prospects in E-commerce. However, the existing E-commerce recommendation system faces many problems such as the poor recommendation quality which is seriously affected by sparse user evaluation of data, the poor scalability of system and single recommendation which could not cover all the interests of users. Aimed at the problems above, the paper did some helpful exploration and research on recommendation technique and architecture of personalized e-commerce recommendation system.In the thesis personalized e-commerce recommendation system was firstly discussed from the point of view as representation way, study content and main methods. Secondly, a variety of intelligent recommendation techniques in recommendation system such as Collaborative Filtering, Information Filtering, Data Mining, Horting Map and Mixed Recommendation were analyzed to summarize the advantages and disadvantages of every recommendation techniques, and then put forwards a collaborative filtering recommendation improved technique based on users interest clustering. Thirdly, the clustering highly similar to objectives was chosen as nearest neighbor to search for objectives by way of clustering the similar items of user dominant interest and user hidden interest. Fifthly, the effectiveness of Recommendation improved technique was validated by experiments, and the results showed that the collaborative filtering recommendation improved technique based on users interest clustering effectively solved the difficult recommendation because of less user rating and gave a better recommendation quality and scalability than the traditional recommendation methods. Finally, the book intelligent recommendation system was designed and developed based on the above recommendation technique.This thesis did some exploration and study about e-commerce recommendation techniques, developed a book intelligent recommendation system which has well recommendation quality and customer satisfaction and gave E-commerce intelligent information service some reference value.