A Novel Interval Grey Number Prediction Model Given Kernel and Grey Number Band

The interval grey number prediction model provides a promising way to handle small example and poor information problems. However, the performance of current interval grey number prediction models still have some defects due to the limitation of accumulation of errors and grey degree enlargement. In this paper, a novel interval grey number prediction model is established through the kernel and area sequence of grey number band The new model does not only ease the impact of extreme data in modeling sequence on model precision through generating average value, but also does not have the defects of accumulation of errors and grey degree enlargement problem. Finally, the novel model is adopted to forecast the amount of ground subsidence. The comparison between the novel model and other models is conducted The results show that the new proposed model is the best among other three traditional models.