This study proposes a framework of using a hidden Markov Chain (HMC) model in analyzing customers’ behaviors from probabilities viewpoints. The philosophy of an HMC model is that the state is changed in accordance with the transition matrix and the observer only sees the output of the random functions associated with their respective state and cannot directly observe the states. Thus, an HMC model is one of the best tools to observe the evolution of systems over repeated trials in successive time periods where the state of the system in any particular period is uncertain. The transition matrix is to compute conditional probabilities of being in a future state given a current state. The emission matrix is to describe the manner where the system makes emission under one state. The transition- emission transformation is used to describe the manner where the system makes changes from one period to the next. A clear and well-defined procedure is summarized. In addition, a briefcase illustration of library users is provided. The major advantage of applying an HMC model is that it is able to explain the possibly heterogeneous nature of sets of observed sequences and to handle partially observed training data.
