[1515 Köne AI, Büke T. Forecasting of CO2 emissions from fuel combustion using trend analysis. Renew Sustain Energy Rev. 2010;14:2906-15.] |
Trend analysis |
A simple model that can be used with the available data |
It is limited to situations of linear behavior |
[1616 Pao HT, Tsai CM. Modeling and forecasting the CO2 emissions, energy consumption, and economic growth in Brazil. Energy [Internet]. 2011;36:2450-8. Available from: http://dx.doi.org/10.1016/j.energy.2011.01.032 http://dx.doi.org/10.1016/j.energy.2011....
] |
Univariate gray prediction model |
Possibility to represent complex systems with limited amounts of data |
The performance of the model in comparison with other models depends on the database used |
[1717 McKitrick R, Strazicich MC, Lee J. Long-term forecasting of global carbon dioxide emissions: Reducing uncertainties using a per capita approach. J Forecast. 2013;32:435-51.] |
Regression for stationary samples, considering the possibility of structural breaks |
The reliability of population growth predictions is high, same advantages as trend analysis |
Limited to stationary situations |
[1212 Réquia WJ, Koutrakis P, Roig HL, Adams MD. Spatiotemporal analysis of traffic emissions in over 5000 municipal districts in Brazil. J Air Waste Manag Assoc [Internet]. 2016;66:1284-1293. Available from: http://dx.doi.org/10.1080/10962247.2016.1221367 http://dx.doi.org/10.1080/10962247.2016....
] |
Top-Down approach for forecasting vehicle emissions in the period between 2001 and 2012 |
Alternative to obtain a more detailed emission distribution profile |
Need a consistent database not only with emissions but also with generating mechanisms |
[1111 Schulz JR da S, Ruppenthal JE. Aplicação Da Metodologia De Box & Jenkins Para Análise Das Emissões De Dióxido De Carbono No Brasil. Reun Rev Adm Contab e Sustentabilidade. 2019;8:1-11.] |
Box & Jenkins, ARIMA (0,1,2), in-sample predictions |
Able to model and predict the behavior of CO2 emissions with non-stationary data. Arima model can validate the results obtained with other models |
The variable, CO2 emissions, is explained only by past values, without taking into account the interference of other factors |
[1818 Acheampong AO, Boateng EB. Modelling carbon emission intensity: Application of artificial neural network. J Clean Prod [Internet]. 2019;225:833-56. Available from: https://doi.org/10.1016/j.jclepro.2019.03.352 https://doi.org/10.1016/j.jclepro.2019.0...
] |
Artificial neural networks: Multilayer Perceptron with back-propagation |
Efficient in dealing with non-linear, complex situations and with diffuse data. It is not necessary to know specific mathematical relationships between variables |
Need to determine the number of neurons through trial and error |
[1919 Wu W, Ma X, Zhang Y, Li W, Wang Y. A novel conformable fractional non-homogeneous grey model for forecasting carbon dioxide emissions of BRICS countries. Sci Total Environ [Internet]. 2020;707:135447. Available from: https://doi.org/10.1016/j.scitotenv.2019.135447 https://doi.org/10.1016/j.scitotenv.2019...
] |
Conformable non-homogeneous gray fractional model |
Possibility to represent complex systems with limited amounts of data |
Low-performance results in cases of random, non-linear, or non-stationary data |
[2020 Ahmed S, Ahmed K, Ismail M. Predictive analysis of CO2 emissions and the role of environmental technology, energy use and economic output: evidence from emerging economies. Air Qual Atmos Heal. 2020;13:1035-44.] |
Gray System Model |
Possibility to represent complex systems with limited amounts of data |
Limited to linear systems, but there are ways to deal with this limitation |