| With the popularization of Internet and the development of E-Commerce, the structure of E-Commerce web site became more and more complex. This situation made it hard for consumers to find the products and services they wanted. To address this issue, recommendation systems were proposed to suggest products and to provide consumers with information to help them decide which products to purchase.Recommendation systems can enhance E-Commerce sales by converting browsers into buyers, increasing cross-sell and building loyalty to prevent user losing. Presently recommendation systems have gradually become an important part in E-Commerce IT technologies, more and more research papers about recommendation systems in E-Commerce appeared in many kinds of conferences and journals.Althrough recommendation systems in E-Commerce have been very successful in both research and practice, challenging research problems remain. Aimed at the main challenges of recommendation systems in E-Commerce, this thesis explored and reseached some key technologies of recommendation systems in E-Commerce. The main research works in the thesis included research of recommendation quality, research of real-time requirement of recommendation, research of recommendation systems based on web mining and research of arthitecture of recommendation systems in E-Commerce.The main research results of this thesis are as follows:1). Item-rating-prediction-based collaborative filtering algoritm. With the expansion of E-Commerce systems, the magnitudes of users and commodities grow rapidly, resulting in the extreme sparsity of user rating data. This situation makes the quality of recommendation systems decreases dramatically. To address this issue, we proposed a collaborative filtering recommendation algoritm based on item rating prediction. This method predicted ratings of un-rated item by the similarity of items, and then the nearest neighbors of target user were calculated with a new similarity measure method. The experiment results suggested that this method could efficiently overcome the extreme sparcity of user rating data and provide better recommendation results than traditional CF algorithms.2). Item-clustering-based collaborative filtering algoritm. In large E-Commerce systems, the real-time requirement of recommendation system is hard to be satisfied. To address this issue, we proposed a collaborative filtering recommendation algorithmbased on item clustering. This method first clustered items by the users' rating on items, based on the similarity between target item and cluster centers, the most similar clusters were selected as the search space to search the nearest neighbor of target item. The experimental results suggested that this method could efficiently improve the real-time response speed of recommendation systems.3). A framework of recommendation system based on web mining. Traditional collaborative filtering recommendation technique is hard to provide recommendation service for unregistered users. To overcome this problem, we suggested a framework of recommendation system based on web mining. This method first clustered web usage data, web content data and web structure data respectively, then provided high-quality recommendation services based on mining results. Compared with traditional collaborative filtering techniques, recommendation systems based on web mining are convenient for users because user needn't to provide user-rating data explicitly.4). An architecture of recommendation systems with multiple recommendation system. Traditional recommendation system is a sole tool with only one recommendation model. Architecture of recommendation system with multiple recommendation models we proposed in this thesis could provide multiple recommendation models, all recommendation models were managed in the same way. This method could efficiently meet different recommendation requirements in different situation of E-Commerce web site. Based on the architecture proposed, we designed and implemented a prototype called...
|
PDF Full Text Request