Abstract
Because broadcasting technologies and networks have been continuously upgraded, we are now able to access extremely large numbers of channels and TV programs in daily life. However, it has become difficult for users to find interesting TV programs from among the multitude of programs available. In this paper, we propose a TV-program recommendation system that utilizes the user's life log data such as browsing histories on WWW, locations visited, personal scheduler on PC, and TV-programs watched. We implement prototype to evaluate effectiveness of our approach. A subjective experiment shows that life log data is effective in improving the quality of program recommendations.
