Multivariate discrete grey model base on dummy drivers
Purpose – The purpose of this paper is to solve the problem that the qualitative relative factors cannot be employed in traditional multivariate grey models. Design/methodology/approach – First, a new model is constructed though introducing dummy drivers. Then, the parameters estimation method and recursive function of the model are discussed. Furthermore, dummy driver setting, pre and post test methods of dummy drivers are proposed. At last, the per capita income forecasting of rural residents in Henan province of China is solved with the proposed model. Findings – The proposed model is the reasonable extension of original one. The accuracy of it is higher than former model. In the case study, the forecasting results of proposed model are compared with other grey forecasting models, and prove that proposed model has not only high accuracy, but also clear physical meaning. Practical implications – The method proposed in the paper could be used in policy effect measure, marketing forecasting, etc., when the predictor variables are influenced by some qualitative variables. Originality/value – It will promote the accuracy of multivariate grey forecasting model.