New grey prediction model and its application in forecasting land subsidence in coal mine

Mining subsidence destroys environment seriously and is difficult to forecast because the parameters in prediction model are difficult to obtain. As there are many uncertainties in mining subsidence, we forecast it by grey prediction model. Traditional GM (1,1) model predict for a time series. In this paper, the panel data are studied and are viewed as a sequence in which elements are matrix based on cross-sectional data, and the mean sequence of row vector GM (1,1) model, mean sequence of column vector GM (1,1) model and the cell volume sequence GM (1,1) model are established, respectively. Combining these grey models, we build prediction model of cross-sectional data matrix sequence. Thus, the scope of grey prediction has been expanded, and grey forecasting theory has been enriched. Using the newly built predictive models, we study the land deformation due to mining of Pingdingshan coal mine in Henan Province. Practical verification and model accuracy test show that the grey model can make accurate predictions, with a good agreement between the predictive value and actual value. It can provide effective and accurate information and also can provide an important reference for the reclamation planning of surface environment.