Method for Inferring High-Order and Dynamic User-Context

We propose an inference system that determines user context. In particular, we focus on high-order and dynamic information such as userpsilas mood and actions, which cannot be measured via any single sensor, and whose inference logics are individual for each user. Therefore, we propose an inference system, which enables us to easily aggregate all sensing information based on a Bayesian belief network and helps general users make their own networks by reducing the technicality of a method for constructing the networks.