基于LS-SVR、BP-ANN和MLR模型的PM_(10)浓度预测

Prediction of PM_(10) Concentrations Based on LS-SVR,BP-ANN and MLR Models

利用宁东能源化工基地PM10和气象监测数据,分别采用LS-SVR、BP-ANN和传统MLR模型预测PM10浓度变化。结果表明,较BP-ANN模型、MLR模型,LS-SVR模型能更好地刻画PM10浓度与各气象因素间的非线性相依关系,更准确地预测PM10浓度。

Using ambient PM10 concentrations and meteorological data of Ningdong Energy and Chemistry Industry Base, predicted PMIo concentrations variation based on LS-SVR, BP-ANN and traditional MLR models, respectively. It was shown that the LS- SVR model could better depict the nonlinear dependency relationship between PM10 concentrations and meteorological factors, more accurately predict PM10 concentrations, comparing to BP-ANN and MLR.