摘要
本文基于MATLAB6.5平台编程,运用非线性BP神经网络对我国外汇储备规模进行预测分析。以我国历年外汇储备数据为训练样本,进行网络训练与检验。结果表明,BP神经网络具有良好的预测性能。同时发现了我国外汇储备存在超常规增长,并在文章最后提出了相关的政策和建议。
Based on MATLAB6.5 programming, BP neural network of foreign exchange reserves forecast in China was construtted. According to former foreign exchange reserves data which were taken as training data are normalized, the networks respectively using non-normalized training data and normalized were trained and verified, the results shown BP network whose training data was normalized has better forecast capability. The speed of China＇s foreign exchange reserves growth was very high.
出处
《未来与发展》
CSSCI
2009年第3期45-48,共4页
Future and Development
关键词
外汇储备
BP神经网络
预测
foreign exchange reserves
BP neural network
forecast
