The Application Of Combination Models In Recommendation System Of E-commerce

With the popularization of Internet and the development of E-Commerce techniques, the E-Commerce systems can serve consumers in many ways. But at the same time, its structure has been becoming more and more complex. It is getting hard now for consumers to find the products what they really want. They often last their ways when searching for products, take a long time and find nothing at the end. By interacting with consumers and simulating the behavior of a salesman to suggest products for consumers, a recommendation system of E-Commerce could help the consumers more easier to find the products they really wand to purchase. In today's warm marketing competition, the recommendation systems of E-Commerce could increase the marketing quotient of a network shop by attracting and retaining consumers with its easy ways in finding and purchasing goods.Although the recommendation systems of E-Commerce have been developed very successful in both research and practice and the scale of them have been extended, challenging problems in research still remain today. Aim at the main challenge of product recommendation, the paper has explored and researched some key technologies on the design about the recommendation algorithm and the structure of a recommendation system of E–Commerce. It has carried on a beneficial exploration on research of a recommendation system of E-Commerce based on the combination recommendation pattern. its main research works can be concluded as bellow:(1) The paper has investigated various mainstream recommender systems and analyzed their strengths and weaknesses.(2) The paper has proposed a new model for the Recommendation System of E-commerce based on the combined recommendation model. Its main idea is that it uses multi-models recommendation algorithms to provide a good recommendation of products for customers. With this architecture, weaknesses could be overcome and advantages could be developed.(3) A research on how to capture and then build a user profile has been done in the paper. A new method to build a user profile based on both the user's explicit interests and implicit interests has been presented by analyzing their gain and expression ways.(4) The paper has also researched the traditional item-similar algorithm and the user-similar algorithm. New computation strategies have been proposed and the...