摘要
简要介绍了北京市空气质量动态统计预报系统的构成、特点,以及所采用的线性回归模型LRM,分类判别树CART模型,CART与LRM结合的模型,动态统计预报模型DSM,多点预报模型MPDSM5种预报模型,分析了不同预报模型的特点和性能。所建立的动态统计预报模型DSM有良好的预报性能,减小了在高污染季节的预报误差;多点空气质量动态预报模型也具有较好的预报性能。
The composition and characterization of the dynamic air quality forecasting system in Beijing were introduced.Five models,including linear regression model(LRM),classification and regression tree(CART) model,hybrid model of CART and LRM,dynamic statistic model(DSM) and multi-points air quality dynamic statistic model(MPDSM),were also introduced,with characterization and performance of these models analyzed.The DSM model has excellent performance in air quality prediction and the prediction error has been decreased obviously,while the MPDSM model has good prediction ability too.
出处
《环境科学研究》
EI
CAS
CSCD
北大核心
2004年第1期70-73,共4页
Research of Environmental Sciences
关键词
污染预报
动态统计
预报模型
air quality forecasting
dynamic statisic
forecast model