Unequal interval grey model based on dynamic correction of time-distance weight

The predicted results of the traditional unequal interval GM (1,1) model always have greater residuals, and a large variety of residual rates, which needs to establish a residual error model to amend them. In view of this situation, a grey prediction model for slope displacement based on the correction of time-distance weight is established by redistributing the time-distance weight in the process of data accumulation or regression, and by determining the optimal time-distance weights after research and checking. During the data regression and reduction, time-distance weights are corrected dynamically based on the trends in the residual rate. The values of dynamic correction make predictions more close to the monitoring results. In the model, both time-varying and gray property are adequately considered to reduce the whole prediction error and improve the prediction accuracy. The case study shows that the fitting precision is high and the prediction is reliable. The proposed model, which is of a certain theoretical and practical significance, can be employed to predict effectively the tendency and results of slope displacement in the short term and middle term.