基于IG-LASSO模型的城市空气质量指数混合预测研究

Hybrid Forecasting of Urban Air Quality Index Based on IG-LASS Model

空气质量指数是各个地区空气污染状况的数据表征,可用于政府对城市空气污染的控制。论文使用天津市2014年1月1日-2016年4月30日的空气质量数据和气象数据,建立一个基于IG(信息增益)和LASSO(最小绝对收缩率和选择算子)的空气质量指数混合预测模型,对未来一天的空气质量指数进行预测。整体实验由预测模型选取、特征变量选取和混合预测3个部分组成。实验结果说明基于IG和LASSO的空气质量指数混合预测模型要比单独使用LASSO模型的预测准确性要好,其误差率为4.75%,并且空气质量指数混合预测模型也可以有效的减少输入变量的数量以及降低模型的复杂程度。同时,也得出天津市空气质量指数的预测准确度受PM_(10)、PM_(2.5)、NO_2和SO_(2)4种空气污染物浓度影响较大,与风向、天气现象和风力关联性不强的结论。

The air quality index can be used by the government agency to characterize the quality of the air at different locations, and based on which to manage and control local and regional air quality. The main objective of the present study is to build a mixed prediction model of air quality index based on information gain and LASSO by use of Tianjin air quality data and meteorological data of the period from January 1, 2014 to April 30, 2016. Then the model is used to predict the air quality index in the next day. The whole experiment was composed of forecast model selecting, feature selecting and variable hybrid forecasting; and it proved that hybrid forecast model based on information gain and LASSO was performing better than the sole LASSO model, its MAPE is 4.75%. And it also could cut down the number of input variables and reduce the complexity of the model. Impacts of the concentrations of four air pollutants, i.e. PM（10）, PM（2.5）, NO2 and SO2, on Tianjin＇s air quality index was great, but not the impacts of wind direction, wind force and weather phenomenon.