ABSTRACT
Large-scale data has been generated in the real daily life by many kinds of electric devices. We can utilize such data to construct computational human models that predict user's behavior, consumer's preference, and so on. Probabilistic models, statistical learning and probabilistic inference method can be applied for computational human modeling. In this paper, examples of applications using Bayesian networks are shown. Bayesian networks have advantage that can represent mutual interaction for situation depend user's preference models. One good example is a contents recommendation depends on different situations and users. In order to get large-scale data corresponding to many kinds of different situations, point of act data collection scheme during service providing is necessary. We call this data collection scheme 'research as a service'. This human modeling framework is also discussed as a key technology for service engineering to improve productivity of service industry.
