Classification of customer loyalty based on Hidden Markov Model

In recent years, many companies have given customer loyalty a high priority on their list of business needs because customer loyalty is essential to their success. Companies must recognise the loyalty and characteristics (price-driven, service-driven or quality-driven et al.) of their customers and market to them appropriately. In this paper, we put forward a customer loyalty analysis process based on customer repurchase, customer price perception, service perception and quality perception. In the data mining process, we analyse the prior customer loyalty information, mine the potential customer information and predict the customer's future purchase. We give a method to set up a statistical model for transition probability matrix of purchase proportion. According to Bayesian rule, we obtain the conditional probability, and calculate the equation that is referred as the likelihood function, and then design a classifier based on the Hidden Markov Model (HMM) for discovering which customer is loyal and which is not loyal.