Forecasting CO2 emissions from Chinese marine fleets using multivariable trend interaction grey model

A reliable prediction of CO2 emissions from marine fleets plays an important role in the low carbon development of shipping industry. However, CO2 emissions from marine fleets have its own inherent trends and the influencing factors might have interaction effects, and these problems make it difficult to build an accurate prediction model. To this end, a novel multivariable trend interaction grey model, named TIGM(1, N), is proposed in this paper. TIGM(1, N) extends the grey prediction model by integrating three different terms, i.e., interaction, trends, and constant terms into the grey action terms of the classical multivariable grey model. Compared with the classical multivariable grey model, TIGM(1, N) effectively reflects the impact of input variable interactions and trends on the system's behavior. To increase the accuracy, the new model's adjustment coefficient is optimized to obtain optimal time-response function values. The experimental results show that TIGM(1, N) outperforms linear regression models and other variants of the grey prediction models in predicting the CO2 emissions from Chinese marine fleets. Finally, the new model is applied to predict the marine fleets' CO2 emissions during 2016-2018 and the results demonstrates the feasibility of the proposed model in low carbon development plan of shipping industry and its value in formulating environmental policies. (c) 2021 Elsevier B.V. All rights reserved.