Abstract
Recently, it is possible to measure and observe human behavior in everyday life due to development of sensor and information processing technology. Furthermore, data-driven modeling of daily human behavior is becoming possible using a large number of measured real data. One of realistic approaches to developing an everyday life behavior simulator is to integrate multiple behavior models that are constructed from different real data in a complementary style. In this paper, the authors propose a method for integrating multiple probabilistic models using everyday life terminology.
