A novel varistructure grey forecasting model with speed adaptation and its application

Traditional grey forecasting models cannot achieve satisfactory predictive accuracy on approximate non-homogeneous exponential sequence with acceleration, velocity and constant perturbation terms. Hence, it is hard to forecast the developing trend of non-homogeneous exponential data sequences widespread in the real world. It is for this reason that an extended intelligent grey forecasting model with variable speed and adaptive structure (VSSGM for short) is provided in this paper. The proposed model possesses three advantages, namely, strong adaptability to perturbation sequence, self-changing structure, and outstanding accuracy in simulation and prediction. One prominent property is that it can simulate/forecast any given inhomogeneous sequence with acceleration, velocity and constant disturbances, which are the most common perturbations widespread in nature and society. And the proposed VSSGM model can be reducible to DGM(1,1), NDGM, SAIGM and VCGM models in theory. Therefore, the proposed model outperforms the traditional grey models. In order to verify the effectiveness, universality, practicality and feasibility of the model, we conducted an experiment to analyze the precision of the model and employ it to simulate China's industrial added value from 2006 to 2015, then utilize it to forecast the industrial added values from 2019 to 2022. Comparing with the real statistic data, it shows that the proposed model has better prediction performance than traditional grey models. (C) 2020 Published by Elsevier B.V. on behalf of International Association for Mathematics and Computers in Simulation (IMACS).