摘要
在目前,汇率预测的方法一般采用人工神经网络和贝叶斯估计方法,但这些方法一般都是基于汇率数据本身进行的。然而汇率的变动实际上是由政治、经济、心理等因素造成的。因此,利用这些新闻消息应该更具有可行性。该文讨论了如何根据新闻消息利用贝叶斯网络来进行汇率的趋势预测,分析了如何才能改进预测精度。同时还实现了一个寻找全局最优贝叶斯网络的选择算法。
At present ,generally methods of forecasting the foreign currency rate are Artificial Neural Network and Em-pirical Bayes,but they are based on the data of the foreign currency rate.However the change of the foreign currency rate is led by political,economic and mental factors.So forecasting with the economic news should be more reasonable.This paper discusses how to forecast the trend of foreign currency rate with Bayesian network based on the economic news and how to improve the accuracy.And an algorithm of searching the best Bayesian network in the global space is implemented.
出处
《计算机工程与应用》
CSCD
北大核心
2002年第15期250-252,256,共4页
Computer Engineering and Applications
