Detecting non-habitual life behaviour using probabilistic finite automata behaviour model

This paper proposes modeling of human behavior and a method of finding irregular human behavior. Initial experimental results demonstrated usefulness of proposed model and method. Firstly, by analyzing human behavior based on observation of human habitual life, it is shown that human behavior can be regarded as probabilistic finite automata. Then a method of constructing human behavior knowledge and an algorithm to distinguish abnormal behavior different from habitual life behavior are proposed based on Hidden Markov Model. Finally, significance of the proposed method is described by some experimental results.