【作者】 许琪； 曹振宇； 周根贵；

【Author】 Qi Xu~2,Zhenyu Cao~1,Gengui Zhou~1 1 College of Business Administration,Zhejiang University of Technology,Hangzhou,310032) 2 Hangzhou branch,Citibank China Co.,Ltd.,Hangzhou, 310003)

【机构】 浙江工业大学经贸管理学院； 花旗银行中国杭州分行；

【摘要】 ATM机的现金管理是每家银行都必须面对的问题。它的目标就是在ATM机中放置最佳数量的现金以满足不确定的需求。有效的现金管理和控制始于先进的自动解决方案,它能够使用先进的算法精确地预测现金的需求。本文回顾已有的ATM机现金需求预测方法,并将它们分为四类:历史经验法、回归分析法、时间序列法和人工智能方法。其中,人工智能方法中的人工神经网络(ANN)和支持向量机(SVM)是目前普遍认可的两种金融时间序列预测方法,得到专家和学者越来越多的关注。更多还原

【Abstract】 Cash management in automatic teller machines(ATMs) is a well-known problem faced by any bank.The purpose is to decide the optimum amount of money that will be placed in the ATMs to satisfy an uncertain demand.Effective cash management and control starts with an automated solution that uses advanced algorithms to accurately predict cash demand,allowing banks to forecast demand and proactively manage cash throughout their network.This paper surveys the existing approaches to cash forecasting and subdivides them into four categories:historical experience method,regression analysis method,time series analysis and artificial intelligence.Special attention is paid to two methods of artificial intelligence-ANN and SVM,which have increasingly popular in finance forecasting.更多还原