Constructing a predicting model for ecological footprint in G20 countries through artificial neural network

Abstract

The purpose of this research is to build an estimated model for the ecological footprint (EF) in the G20 countries in the period of 1999–2018, the last two decades. These countries have faced extensive EF changes and developments. This aim, indices (The Global Footprint Network and World Bank) has been used. Artificial neural network (ANN) models have been used for data analysis and index fitting. The findings of this research showed that the EF has been increasing in China with an average of (16.37), France (14.58), Brazil (9.37), and the United States (3.66). Root mean square error value in the first model (0.140), in the second model (0.0275), in the third model (0.0275) in the fourth model, is equal to (0.0608) and in the fifth model it is equal (0.11484). Therefore, the accuracy of EF prediction in neural network models 2 and 4 is 97.5% and 97.5%, respectively. Also, we believe that in better management of the EF, the use the ANN can be efficient and effective in forecasting the G20 countries. This study can provide two important contributions to the managers of this country in examining the environmental degradation of the G20 countries. First, the use of ANN is one of the methods of artificial intelligence that has a very high and acceptable accuracy in selecting research variables. Second, we used ANN models to solve the problem of prediction accuracy and selection of effective variables.