摘要
通过对神经网络应用实例的论证 ,探讨B -P神经网络应用于环境空气质量预报的可行性及其先进性。B -P神经网络对污染物浓度变化趋势的预测较线性回归统计模式更为敏感 ,其预测值的绝对误差也比线性回归统计模式小。显示出其逼近精度高、训练学习速度快、对基础数据时间长度要求不高的优越特性。
Through discussing application of B-P Nerve Network, to inquire the feasibility and advancement of B-P nerve network applying in environment air quality forecast. The B-P nerve network forecasts trend of the pollutant's concentration changing is available than the Linest, the absolute error of its forecast results is small than the Linest.
出处
《云南环境科学》
2003年第B03期52-54,共3页
Yunnan Environmental Science