A Novel Combined Model for Air Quality Index Forecasting in Changchun

With the rapid development of economy and continuous improvement in people’s living standards, predictions air quality index have attracted wide attention. In this paper, a new feature selection method (Pearson-MI) combined model construction (modified inverse variance method) were proposed to study (AQI) its influencing factors Changchun. The Pearson-MI selects that affect AQI Changchun City from many factors. This reduces RMSE LSTM XGBoost by 27% 5% MAE 41% 5%, respectively. A combines XGBoost, SVR, RF, was constructed using predict City. modified resulted 2% reduction 0.6% compared with unmodified model. numerical results our show prediction accuracy is obviously higher than basic model, further improved under selection.