An improved grey multivariable time-delay prediction model with application to the value of high-tech industry. (1st March 2023)

• Record Type: Journal Article
• Authors: Zhou, Huimin; Dang, Yaoguo; Yang, Deling; Wang, Junjie; Yang, Yingjie
• Is Part Of: Expert systems with applications. Volume 213:Part B(2023)
• Publication Date: 2023-03-01
• Subjects: Grey multivariable model -- Forecasting -- Time-delay -- Linear correction term -- High-tech industry; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2022.119061
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: A novel discrete grey multivariable time-delay model is proposed. A linear correction term is designed to broaden the scope of the proposed model. The parameters for representing the time lag effect are developed in the new model. The proposed model outperforms the competitive models. Abstract: To analyse the time lag effects between independent variables and dependent variables, we propose a discrete time-delay grey multivariable model DMTGM 1, N τ . There are three improvements in this new model compared to the existing models. First, the time lag parameters are assigned different values for each independent variable. A linear correction term expands the new model. Second, with the given time lag, the least square method can be used to calculate the parameter vector. The time response function of DMTGM 1, N τ is generated, which has the advantage of eliminating the jumping errors between discrete and continuous functions over the existing grey forecasting models. Third, all of the feasible combinations of the time lag parameters are compared by using a traversal algorithm to identify the best values with the minimized mean absolute percentage error (MAPE). In three different case studies, the performance of the new model is evaluated and compared to that of a number of mainstream grey models as well as non-grey models. According to the findings, the newly designed model performs significantly better than the compared models.