An author recommendation system using both content-based and collaborative filtering methods.

摘要

The immense growth of the World Wide Web has led to the overload of information available online, making it difficult for users to process all the information. Recommender systems are a popular technology that attempt to overcome this problem by exploiting the information provided by users on the available items to predict new items for the user. In this thesis we use social tagging information present in a social reference manager and the textual information to make recommendations.;We propose a hybrid recommender algorithm that combines two of the most popular algorithms, namely, collaborative filtering (CF) and content-based (CB) to make author recommendations. Experiments demonstrate that our hybrid algorithm improves the quality of the recommendations and solves the inherent problems of the two approaches.