A novel multivariate time-lag discrete grey model based on action time and intensities for predicting the productions in food industry

The food industry is a pillar industry and plays an important role in the development of China's national economy. In view of the complex environments and the characteristics of fluctuating development in which the Chinese food industry are located, this paper delves into the mechanism of action of relevant factors on the system behavior and proposes a grey prediction model based on action time and intensities considering the widespread existence of time-lag effects that is applicable to the prediction of the food industry. Firstly, considering the action time and intensity of the factors, this paper introduces the action duration, the parameters of time lag and intensity coefficient to expand the multivariate discrete grey prediction model. Secondly, the bidirectional grey A-S (Area-Slope) comprehensive relational model is used to calculate the degree of correlation between relevant factors acting on time-effect sub-sequence and the system behavior sequence under different action durations and time lag, to identify the driving factors and estimate their action durations and time lag. Then, the driving factors are incorporated into the grey prediction model based on action time and intensities, and the structural parameters and intensity coefficients of the model are estimated by using the least square and particle swarm algorithm. Finally, by fitting and predicting two examples in the food industry, the results of the proposed novel model are compared with those of six other prediction models to verify effectiveness. The results demonstrates that the prediction accuracy of the proposed model is better than those of other grey models in time-lag data sequences.