论文部分内容阅读
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 selection of basis functions and the solution of coefficients are two key problems in function approximation. At present, the widely used regression analysis or other mapping methods have limitations. The BP network in artificial neural network is a good solution to these two problems. The theoretical and practical data are compared and verified. The results show that BP network can help to improve the reliability of data processing and prediction accuracy.