| Commercial bank’s capital includes equity capital and debt capital. Demanddeposits is the main part of the debt capital, is also the main source of business capital，represents about40%。. With the faster pace of bank capital flows, the capital forecastbecomes particularly important. Along with the development of information technologyhas brought a sharp increase in the amount of data, data analysis and data miningreaches a new level in the practical application, the accuracy of debt capital forecastbecomes higher.Through the traditional method to estimate debt capital is a rough estimate, it didnot pass the differential analysis of the different categories of customers. Anotherprediction method is established time series model for each customer account, and thencalculates the value of the customer account balance, and then summing; this methodtoo costly, and because of its randomness and unpredictability leads to the decline in theavailability of the model. This topic according to this problem, propose a new predictionmethod and customer segmentation model. Customer segmentation model contains thestable customer segmentation model and segmentation model of customer accounttransactions habits.This topic is the study of accurate description of the bank’s current savings debtcapital. Accurate prediction of demand load assets can bring a better quality ofdecision-making and more profit to the bank. In this topic, includes customersegmentation, trading behavior segmentation and time series analysis. These analyzesalso give banks make better decisions in other areas or products, Customersegmentation model is also a high value in the customer relationship management. Its main methods include K-mean method and Twostep method and Decision Tree andSimple seasonal model of the time series. |
