A novel approach to predict CO2 emission in the agriculture sector of Iran based on Inclusive Multiple Model

Due to the significant effects of CO2 emissions on climate change and global warming, as well as its serious hazards to human health, the prediction of CO2 emission is a vital issue. The main aim of this paper is to evaluate the power of the Inclusive Multiple Model (IMM) as a novel approach to predict CO2 emission in the agriculture sector of Iran. For the same, we implemented the environmental Kuznets curve (EKC) specification and data from 2003 to 2017 for 28 provinces of Iran. In the first level, various specifications were implemented for each of the Multiple Regression (MLR), Gaussian Process Regression (GPR), and Artificial Neural Network (ANN) models. In the second level, an IMM model was implemented for treating the outputs of the best specification out of the MLR, GPR, and ANN models as inputs to an ANN model. The performance of the models was compared with the Taylor diagram and innovation and unique graphs. Findings indicated that the IMM model with C C = 0.81 , RMSE = 0.69, the highest residuals between −5 and 5 (37.84%), and the lowest distance from observation points (1.857) estimated CO2 emission values more precisely. These improvements indicate that there are possible directions for future research activities. Due to the most accuracy of the IMM, it is recommended to use this method to predict CO2 emission to adopt appropriate policies for reducing air pollution.