A Generic Framework for Collaborative Recommendation in a Scientific Network

Social recommender systems (SRS) become more and more popular as they allow use more information about users and alleviate cold-start problem. Traditional recommender systems have concentrated on information or document retrieval and then were extended with methods for music or movie recommendations. SRS are useful especially in the context of online stores: user can quickly find relevant items and a company can improve its income by recommending additional items. A slightly different situation is in the case of research paper recommendation. Usually, researchers are looking for relevant papers using some search engines. Using a few keywords, one can obtain a list of similar works. Let us consider a situation when a researcher likes to find conference or journal where it is worth to publish. He or she receives many "calls for papers" but most of them are not correlated with their researcher interests. In this paper we propose a method for journal or conference recommendation based on social community of researchers. We present a method that combines content-based, collaborative and social approaches to find the best place to publish.