Group recommendation system is an important research area, because there are many situations where a group user take an item such as watching the movie, listening the music and watching the TV contents with their family or friends. In order to suggest group recommendation, understanding of group user and domain is necessary. However, there are few analysis of group recommendation system using real-world dataset, because most researches use synthetic dataset. In this paper, we apply the various methods to real-world dataset, and provide some guides for group recommendation system.