應用類神經網路於自動櫃員機提款預測之研究

Abstract

本研究的目的在於利用倒傳遞類神經網路建立自動櫃員機提款預測模型，提供企業較正確的自動櫃員機提款預測，以便擬定較佳的自動櫃員機補鈔計畫。使該企業在自動櫃員機現鈔的管理達到最佳化，並且有效的降低成本。

由於類神經網路模型同時考慮了線性與非線性關係，且其具有高速計算能力、高容記憶能力、學習能力、及容錯能力等優點。因此本研究藉由類神經網路中的倒傳遞類神經網路建立自動櫃員機提款預測模型。首先，建立模型輸出層和輸入層定義之後，再根據倒傳遞類神經網路學習步驟將預測模型建立起來。本研究所使用的類神經網路軟體為Qnet2000。經由本研究所收集的資料，實際驗證結果顯示：以倒傳遞類神經網路建立自動櫃員機提款預測模型，確實可做為預測自動櫃員機提款的有效方法之一。

The purpose of this research is to construct a forecasting model for cash withdrawals from ATM by using Artificial Neural Network (ANN). In order to provide the better cash withdrawals service, we need a more accurate forecasting model for cash withdrawals from ATM. It will help the company to enhance the cash management and reduce the cost. ANN has error tolerance ability, learning ability, high-speed computational ability, and high-volume memorizing ability. It also considers both linear and nonlinear relationship at the same time. This research aims to construct a forecasting model for cash withdrawals from ATM by using software Qnet2000 that is designed for ANN with Back-Propagation algorithm. First, we define the input and output nodes and set up the parameters and follow the training procedure of Back-Propagation Neural Network (BPN). Then, we construct a forecasting model of cash withdrawals from ATM by using the BPN. A real case was presented to demonstrate the methodology and the results revealed that the BPN could provide an acceptable accuracy for the forecasting model of cash withdrawals from ATM.