A novel q-GM (1,1) modeling with application to greenhouse gas emissions

Purpose This paper focuses on forecasting greenhouse gas emissions as a critical appraisal tool for global warming, climate change, environmental policy and economic planning. There is a need to forecast greenhouse gas emissions for developing effective environmental policies. Therefore, this paper proposes a robust grey prediction model. Design/methodology/approach A novel q-GM (1,1) modeling approach provides prediction of greenhouse gas emissions in Türkiye using the 1990–2022 data set. First, we partition this data set for training (1990–2018) and test (2019–2022) purposes. The performance evaluation of the novel q-GM (1,1) is applied to mean absolute percentage error (MAPE), root mean square error (RMSE), and posterior variance test C and p values. We provide sensitivity analysis with different q-values and data partitions. The performance of the proposed grey model provides robust results in the standard GM (1,1) and discrete grey model (DGM). Findings The proposed q- GM (1,1) model has satisfactory results from the training dataset covering 1990–2018 and the testing dataset covering 2019–2022 based on MAPE, RMSE, C and p values, respectively. We present the forecasted values from 2023 to 2030. These results show that the greenhouse gas emissions will rise in next eight years in Türkiye. Environmental policies should be evaluated by these values. Originality/value This study proposes a novel q-GM (1,1) model for predicted values of greenhouse gas emissions in Türkiye.