Hybrid Recommendation Algorithm Based on Social Interaction and Tag Weight

摘要

In the information age, the recommendation algorithm provides users with an advanced and effective way to search information. However, problems including data sparsity and cold start still exist in the recommended system. In this paper, we employed a hybrid recommendation algorithm based on social interaction and tag weight through analyzing data and experimenting on the Delicious Dataset. Firstly, the data matrix abstracted from the original dataset is divided into multiple groups through users' social relationship so as to alleviate the data sparsity problem, and then we choose the alternative recommendation set by the similarity calculation combined with the weight of the tags. At the same time, if the number of recommended data is less than what is required, the data is supplemented by the items-based and users-based collaborative filtering algorithm. Finally, time parameter is used as an adjusting factor to filter recommendation set. The results demonstrated that the accuracy of experiments has been improved, and the data sparsity and cold start problems have been relieved.