融合ARIMA-LSTM模型的大连市空气质量预测

Air Quality Prediction of Dalian City Using ARIMA-LSTM Model

由于空气质量指数(AQI)时间序列的线性与非线性特征,传统的ARIMA模型对时间序列的建模普遍呈现出一定的局限性,该方法存在着参数选取困难、计算量大等问题,导致模型拟合效果不佳。针对这一情况,本文提出了一种基于ARIMA-LSTM融合模型的空气质量预测方法,将ARIMA模型用于时间序列预测,利用LSTM模型对ARIMA模型预测的误差序列进行校正,最后将ARIMA模型预测结果与LSTM校正的残差序列进行结合,获得最终预测结果。实验结果表明,混合模型预测精度高于单一模型,且融合模型的稳定性和精确度得到进一步改善。

Due to the linear and nonlinear characteristics of the air quality index(AQI)time series,the traditional ARIMA model for modeling the time series generally presents certain limitations,and the method suffers from the problems of difficult pa⁃rameter selection and large computational effort,resulting in poor model fitting.To address this situation,this paper proposes a com⁃bined ARIMA-LSTM model-based air quality prediction method,in which the ARIMA model is used for time series prediction,the LSTM model is used to correct the error series predicted by the ARIMA model,and finally the ARIMA model prediction results are combined with the residual series corrected by the LSTM to obtain the final prediction results.Experimental results show that the prediction accuracy of the hybrid model is higher than that of the single model,and the stability is higher。