Grey Exponential Smoothing Model And Its Application On Air Quality Forecasting

Grey prediction theory has developed rapidly since it was put forward.It has been widely used in many fields.The grey accumulation generating operator is a unique way of grey prediction theory.On one hand,it can clearly show the regularity contained in the original data and smooth the irregularity of the original data.On the other hand,it can provide intermediate information for modeling and help people build more appropriate models to solve problems.In addition,fractional order accumulation generating operator can reduce the perturbation of grey model solution and improve the prediction accuracy.Therefore,the grey fractional accumulation generating operator is introduced into the exponential smoothing model to propose the grey exponential smoothing model.The prediction accuracy of the grey exponential smoothing model and the traditional exponential smoothing model is compared to show that the grey exponential smoothing model has better prediction performance.Firstly,this paper analyses the grey prediction theory and the research status of air quality prediction.The grey accumulation generating operator,time series memory,and the related theory of exponential smoothing are expounded in this paper.Secondly,in order to analyze the influence of the grey accumulation generating operator on the memory of the sequence,the generalized Hurst exponent method is used to calculate the Hurst exponent of the first-order accumulation generating sequence and the original sequence.Then,the grey fractional accumulation generating operator is introduced into the traditional exponential smoothing model to propose the grey exponential smoothing model.The Holt-Winters model,a branch of cubic exponential smoothing model,is suitable for data with seasonal variation characteristics.The optimal solution is found in the parameter space by grey wolf optimization algorithm to improve the prediction accuracy.Finally,the grey Holt-Winters model is applied to the prediction of four cities air quality index in Beijing-Tianjin-Hebei region,which further illustrates the practical significance of the grey exponential smoothing model.