In order to improve people's lifestyles to prevent lifestyle-related diseases, it is important to understand not only the daily schedule but also the acceptability of interventions, such as taking dinner 30 minutes early. In this paper, we propose a new nonnegative tensor factorization method that can analyze collected activity logs and acceptability logs. The proposed method adopts convolutional transformation matrix into factorization form; it enable us to extract activity patterns and also to grasp correspondence of appearance time of each pattern. We show the result of the analysis using the proposed method.