Abstract
Following the improvement of internet technology, it changes the transaction types of customers in banking and solves the inconvenience of traditional finance. Internet banking provides opening an account, remittance, and banking transferring on the website or mobile. More people tend to apply for internet banking to deal with some financial activities. However, some customers might apply the account of internet banking because of the sales promotion of bank business. It easily causes that internet banking accounts are useless or little use for customers. Like the occasional remittance or transferring, they will reduce the use of online banking over time. The data storage capacity of bank will be affected by the above circumstances, and the judgment error of non-target customers will also occur in promoting internet banking.
This study aims to apply the mixture use model to find the characteristic of customers in the preference of internet banking. This model includes the willingness part and the use time of internet banking. Since we can’t observe the willingness to use internet banking, the expectation-maximum algorithm is adopted to estimate the interesting parameters of the mixture use model. In addition, this study considered logistic regression model and random forest method to predict the preference of using internet banking. The real case is from an anonymous bank in Taiwan, and we utilize the mixture use model to explore the characteristic of customers in the preference part and the part of use time. The result shows that age, owning a security account and living region are significant variables in the willingness part of using internet banking. These characteristics of customers also affect the use time of internet banking. This result follows the financial market trends.