摘要
基函数选取和系数求解是函数逼近中的两个关键问题,目前广泛应用的回归分析或其它映射方法均有局限性。采用人工神经网络中的BP网较好地解决了这两个问题,对理论和实际的数据进行了比较与验证,结果表明.BP网络有助于提高数据处理的可靠性与预测的精度.
The fundamental function selection and coefficient solution are important for function approximation method. The general regress analysis or other cast methods have some limitations. The back-propagation neutral net (BP) can give a good solution. Comparison of theoretical calculation with the test data shows that BP neutral net can improve the reliability of data processing and the accuracy of prediction.
出处
《石油大学学报（自然科学版）》
CSCD
1993年第6期126-130,共5页
Journal of the University of Petroleum,China(Edition of Natural Science)
关键词
人工神经网络
BP网络
计算机网络
Neutral net
Regress analysis: Back-propagation net: Data processing
