Living in society and interacting with others, going outside one’s house is almost inevitable for healthy life. There is increasing attention to this behavior, including pervasive computing, medical science, etc. There are various factors to affect the daily going-out behavior such as day of the week, the condition of one’s health, and weather. However, we assume that there should be one’s own rhythms of going-out. In this work, we formulate going-out behavior of a single day with a sequence of Bernoulli distribution, and develop a categorization method that does not require the number of categories a priori, employing Dirichlet process mixture. We also develop a method predicting one’s future state of being out or at home from the extracted patterns, for practical applications (e.g. controlling heating, ventilation, and cooling). Experimental results using time histories of going out/coming home (6 subjects, total 827 days) show our advantages over the existing algorithms (Krumm and Brush. Pervasive2011, Scott et al. UbiComp2011).