【摘要】 针对BP神经网络及各类改进的神经网络模型进行预测易出现收敛陷入局部极小点、网络震荡大、预测值可信度不高等问题,本文提出建立AGO-BP神经网络预测模型观点,并建立了实际应用模型。通过两个实例及多次反复验证,结果显示该模型具有较高的预测精度,且可以比较有效地避免上述问题,该模型具有一定的工程应用价值。更多还原

【Abstract】 The thinking of building AGO-BP neural network model and models for the practical application based on preprocessing the measure data to reveal obviously the internal rules as the start were given aiming at exiting issues of BP neural network and various improved neural network to forecast something,including convergence of the local minimal points,the network frequent shocks,poor reliability of forecasting results.The results indicated that the model was proven to be a certain engineering application value through two actual examples.更多还原