FORECASTING DEMAND FOR CASH OF ATM--USING THE NEURAL NETWORK THEORY

摘要

The invention of as relating to demand forecasting methods of automation using neural network theory to train and learn the trading performance of automated equipment such as ATM or CD group, forecast a device-specific future demand equipment, automation equipment to the prediction in more detail by setting the withdrawal or the amount of the input variables and the past which dates to the output stage of the neural network, found the appropriate weights through learning, then enter the data for the desired forecast period in the completed neural network of input and performs a neural network with the necessary forecasting relates to the method of automated equipment using neural network theory, you can find the demand forecast amount.; This is by appropriate allocation of the limited working capital, reduce airborne hyeonsong needed to plan and reduce unnecessary operations through proper forecasting and optimization to increase the number of transactions can maximize operational efficiency. Furthermore, has the advantage of being scalable with respect to predictions, not only office automation equipment or team-specific (Team), point-by-point demand forecasts.