The Personalized Teaching Resources Recommendation System Based On Social Tags And Mixed-mode Design

With the development of the Internet, people in a broad array of network data more and more difficult to find the data of interest. Personalized recommendation system provides a very good idea to solve this problem. It can be dominant or recessive manner to gather information to generate personalized recommendations. Although people have already made a lot of research in terms of personalized recommendation. However, the personalization system is a great lack of forecasting, especially in the prediction of the user model.In web2.0technology, the social Tag system also obtains good results information on the label aspect, and has been successfully applied to many aspects of the network. The socialization of the Tag system not only allows for a more reasonable classification of network data. Its social characteristics, allows us to dig out a lot of potential users of information and users of social relations.In this paper,it provides a hybrid recommendation system of a social Tag. Its core idea is mainly concentrated in the following areas:1,First, the user model and resource model. It is based on users-tags, resource-Tag weight matrix, these weights based on TF-IDF text information extraction techniques.2,Followed by the synonym extraction. In order to make the Tag to be able to more accurately reflect the real information on the data obtained.3,Once again, the mining of social property. In order to dig out the relationship between potential users and user tag system, the article uses the excavation associated with data mining algorithms Apriori its social attributes. The aim is to improve the accuracy of the user model.4,The interest-based approach to personalized recommendation. In order to more accurately predict the user's model, using the clustering algorithm to cluster. Clustering on interest-based approach, will generally achieve more accurate results.5,Finally, the forecast based on user model, using personalized content-based and collaborative filtering-based hybrid technology.With this study, the proposed model of a teaching resource personalized recommendation system, and some of the details of the concrete realization of the recommendation system is provided. a novel system architecture to achieve on the personalized teaching resources recommendation (for example, the user model predictions is the two data mining techniques combined, which makes the user model may be more accurate). Of course, it can also be used in other directions personalized recommendations.