The Optimization of Forecasting ATMs Cash Demand of Iran Banking Network Using LSTM Deep Recursive Neural Network

One of the problems of the banking system is cash demand forecasting for ATMs (Automated Teller Machine). The correct prediction can lead to the profitability of the banking system for the following reasons and it will satisfy the customers of this banking system. Accuracy in this prediction are the main goal of this research. If an ATM faces a shortage of cash, it will face the decline of bank popularity and in turn will have some costs; and the bank will encounter decreasing the customers use of these systems. On the other hand, if the bank faces cash trapping at an ATM, regarding to inflation in Iran, it will have a negative impact on bank profitability. The aim of this study is to predict accurately to eliminate the posed double costs. Since the information related to the amount of cash is daily, each ATM will have a behavior as time series; and also because the aim of this study is to predict the demand for cash forecasting from all of the ATMs, we are facing data from the type of panel. The methods that are used for forecasting ATM cash demand in this research include: Forecasting by statistical method, MLP neural network method and LSTM deep recurrent neural network. We will compare the results of these methods and show that LSTM deep recurrent neural network method has the best accuracy in forecasting.