Research On Personalized Recommendation Method Based On Labeling

In the21st century, the innovation of technology promotes the rapid development of the network, the most important feature is the explosion of information. Early in the era of information explosion, tools like the expert system and classification of directories and search engines that help people to obtain effective information began to appear However, with the increasing amount of information, traditional tools cannot help users to access effective information from it anymore, in order to solve this problem, recommendation system which with a collaborative filtering recommendation technology provides a new method, because of the simplicity and don’t need to care about the content of recommended project, now are widely used in the traditional recommendation system.As one of the most important products of Web2.0, social tagging has been used by many systems. As tags are used to organize and manage the resources of projects by users, so it becomes a bridge connecting users and resources. Furthermore, by analyzing the tag data can be more comprehensively describe the nature of project and better reflect the user’s interests.In this article, based on the study of traditional collaborative filtering recommendation algorithm and tagging systems, we design a tag-based recommender system in E-Learning area, and we proposes a tag-based recommendation method to improve the exist problem of it, which cannot consider the user multiple points of interest to make recommendation. Firstly, dividing the possible interests of target audience by mining the implicit information in the tag data. Secondly, combining these interests with user-based collaborative filtering recommendation algorithm, and respectively create neighbor users that similar to the target users, then, producing recommendation results through the records of neighbor users. Finally merging all recommendation results and get the finally recommendation results.After preprocessing Delicious publicly available data sets through certain rules, compare the proposed algorithm with user-based collaborative filtering recommendation algorithm and by using precision and recall rate as evaluation indicators commonly used in the recommendation area, experiment result shows that the tag-based recommendation method is significantly better than the traditional collaborative filtering method in producing recommendation. At the same time, the thesis draws tag-based recommendation method into the designed E-learning system. And showing a prototype implementation of the system, furthermore, it is feasibility of the method.