【摘要】 鉴于城市大气环境质量受到诸多复杂因素影响,且各因素间存在多重相关性,本文将偏最小二乘(PLS)分析与人工神经网络径向基网络(RBF)耦合,建立偏最小二乘径向基神经网络模型(PLSRBF),应用于贵阳大气环境质量的检验和预测。实例表明:PLSRBF模型可对原多自变量模型进行降维简化,并可有效提取解释变量信息,防止信息丢失,且具有较强的拟合能力。更多还原

【Abstract】 As urban atmosphere environmental quality was influenced by various complicated factors among which there was,multiple correlation,a Partial Least Square Radial Basis Function(PLSRBF) artificial neural network model which was used to test and predict atmosphere environmental quality of Guiyang was proposed by coupling partial least square analysis and RBF neural network.The results showed that the PLSRBF model could simplify the original multiple variables model via reducing its dimension,and extracted information effectively from independent variables avoiding losing information, which was of high fitting ability.更多还原