A new multivariable grey prediction model with structure compatibility

• Published in: Applied Mathematical Modelling Vol. 75; pp. 385 - 397
• Main Authors: Zeng, Bo, Duan, Huiming, Zhou, Yufeng
• Format: Journal Article
• Language: English
• Published: New York Elsevier Inc 01.11.2019; Elsevier BV
• Subjects: Compatibility; Computer simulation; Dependent variables; Grey prediction; Grey System Theory; Model structural parameters; Multivariable prediction model; Performance comparison; Multivariable prediction model; Performance comparison; Model structural parameters; Compatibility; Grey System Theory
• ISSN: 0307-904X; 1088-8691; 0307-904X
• DOI: 10.1016/j.apm.2019.05.044

•A new grey model with dependent variable lag term, linear correction term and random disturbance term is proposed.•The new model is applied to solve the problems of single structure and poor adaptability of traditional grey models.•The new grey model is fully compatible in structure with the traditional mainstream grey prediction models.•The new model can unify the grey prediction model types and reduce the low level and simple repetition of grey models.•The application in three cases verifies that the performance of the model is much better than that of other grey models. A new multivariable grey prediction model was proposed by adding a dependent variable lag term, a linear correction term and a random disturbance term to the traditional GM(1,N) model. It was theoretically proved that the new model can be completely compatible with the mainstream single variable and multivariable grey prediction models by adjusting and changing the model's parameters. To test the performance of the new model, three case studies were performed. The simulation and prediction results of the new model were compared with those of other grey prediction models. Results showed that the new model had evidently superior performance to other grey models, which confirms that the structure design of the new model is more reasonable than those of the other existing grey prediction models.