KBRS — Keyword based recommendation system in social networks

Abstract

Social networks form an important platform for information sharing and interaction among users. The content from social networks can be used to generate recommendations for users in order to help them to choose what they desire. There exist a lot of recommendation methods currently. In this paper, we propose a keyword based recommendation system (KBRS), where the user's preferences are indicated by keywords. Here, we use a user based collaborative filtering (UCF) algorithm to provide recommendations. In order to support a more efficient and scalable execution, KBRS is implemented in Hadoop, using Map Reduce paradigm.