Open AccessMachine Learned Artificial Neural Networks Vs Linear Regression: A Case of Chinese Carbon Emissions
Author(s) -
Muhammad Jawad Sajid
Publication year - 2020
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/495/1/012044
Subject(s) - artificial neural network , linear regression , regression , regression analysis , work (physics) , relation (database) , computer science , mean squared error , econometrics , artificial intelligence , linear model , machine learning , statistics , mathematics , data mining , engineering , mechanical engineering
China is the topmost source of world’s carbon emissions. Keeping this in view, a lot of work has focused on evaluating the relation between the Chinese carbon emissions and its drivers. However, these works mostly employ different types and extensions of the regression model to estimate the relations. The popular machine learning approaches like the artificial neural networks (ANN) are mostly overlooked in this regard. Furthermore, the studies based on the ANN and its different extensions often boast its superiority over the regression analysis. This claim has also not yet analysed for the relationship between a regions carbon emissions and their drivers. This study fills these critical research gaps. The results showed that the linear regression model with lesser ‘mean squared error’ outperformed the ANN model with linear activation code. This study can be a good starting reference for advanced future work on this much neglected research gap.