Investigation of the Relationship Between CO2 Emission and Primary Energy Consumption with Quantile Regression Model

Hayriye Esra Akyüz

Article Details

Investigation of the Relationship Between CO2 Emission and Primary Energy Consumption with Quantile Regression Model

Authors:

Hayriye Esra Akyüz

Publisher:

Duzce University

Main Topics:

Mühendislik

Journal:

Düzce Üniversitesi Bilim ve Teknoloji Dergisi Vol.11 Issue 3, pp.1529-1545

Document Type:

Article

Publication Date:

7/2023

Document Language:

Turkish

Document ID (Org No.):

2001296367

pISSN:

2148-2446

Keywords:

Abstract:

 In this study, it is aimed to model the relationship between CO2 emissions and primary energy consumption between the years 2000-2020 in Turkey, with quantile regression analysis, which can be used in cases where the assumptions of linear regression analysis are not provided. In the first part of the study, after obtaining the regression model, the results were obtained in the presence of outlier observations, and the performance of the quantile regression was evaluated. Root mean square error RMSE and mean absolute error MAE criteria were used to select the appropriate model. According to the results obtained; between the years 2000-2020 in Turkey, the average CO2 emission and primary energy consumption are 289.6268.12 million tonnes and 4.661.21 exajoules, respectively. It has been found that 98.74 of the change in CO2 emission is explained by primary energy consumption and in the presence of heteroskedasticity problem, quantile regression analysis gives more appropriate results than linear regression analysis. According to quantile regression model; it has been determined that one exajoule increase in primary energy consumption will cause an increase of approximately 54.3 million tonnes for τ0.50 quantile of CO2 emissions. In the presence of outliers, the quantile regression models obtained for for τ0.25 and for τ0.50 were the most appropriate models and they gave very close results. It is suggested that these models can be used for inference in modeling the relationship between CO2 emissions and primary energy consumption.

 In this study, it is aimed to model the relationship between CO2 emissions and primary energy consumption between the years 2000-2020 in Turkey, with quantile regression analysis, which can be used in cases where the assumptions of linear regression analysis are not provided. In the first part of the study, after obtaining the regression model, the results were obtained in the presence of outlier observations, and the performance of the quantile regression was evaluated. Root mean square error RMSE and mean absolute error MAE criteria were used to select the appropriate model. According to the results obtained; between the years 2000-2020 in Turkey, the average CO2 emission and primary energy consumption are 289.6268.12 million tonnes and 4.661.21 exajoules, respectively. It has been found that 98.74 of the change in CO2 emission is explained by primary energy consumption and in the presence of heteroskedasticity problem, quantile regression analysis gives more appropriate results than linear regression analysis. According to quantile regression model; it has been determined that one exajoule increase in primary energy consumption will cause an increase of approximately 54.3 million tonnes for τ0.50 quantile of CO2 emissions. In the presence of outliers, the quantile regression models obtained for for τ0.25 and for τ0.50 were the most appropriate models and they gave very close results. It is suggested that these models can be used for inference in modeling the relationship between CO2 emissions and primary energy consumption.

 In this study, it is aimed to model the relationship between CO2 emissions and primary energy consumption between the years 2000-2020 in Turkey, with quantile regression analysis, which can be used in cases where the assumptions of linear regression analysis are not provided. In the first part of the study, after obtaining the regression model, the results were obtained in the presence of outlier observations, and the performance of the quantile regression was evaluated. Root mean square error RMSE and mean absolute error MAE criteria were used to select the appropriate model. According to the results obtained; between the years 2000-2020 in Turkey, the average CO2 emission and primary energy consumption are 289.6268.12 million tonnes and 4.661.21 exajoules, respectively. It has been found that 98.74 of the change in CO2 emission is explained by primary energy consumption and in the presence of heteroskedasticity problem, quantile regression analysis gives more appropriate results than linear regression analysis. According to quantile regression model; it has been determined that one exajoule increase in primary energy consumption will cause an increase of approximately 54.3 million tonnes for τ0.50 quantile of CO2 emissions. In the presence of outliers, the quantile regression models obtained for for τ0.25 and for τ0.50 were the most appropriate models and they gave very close results. It is suggested that these models can be used for inference in modeling the relationship between CO2 emissions and primary energy consumption.

 Bu çalışmada doğrusal regresyon analizinin varsayımlarının sağlanmadığı durumlarda kullanılabilecek olan kantil regresyon analizi ile Türkiyede 2000-2020 yılları arasındaki CO2 emisyonu ve birincil enerji tüketimi arasındaki ilişkinin modellenmesi amaçlanmıştır. Çalımanın ilk böümünde uygun regresyon modeli elde edildikten sonra, veri setine aykırı değerler eklenerek bu gözlemlerin varlığında sonuçlar yeniden elde edilmiş ve kantil regresyonun performansı değerlendirilmiştir. Uygun modelin seçiminde hata kareler ortalamasının karekökü RMSE ve ortalama mutlak hata MAE kriterleri kullanılmıştır. Elde edilen sonuçlara göre; Türkiyede 2000-2020 yılları arasında ortalama CO2 emisyonu 289.6268.12 milyon ton ve ortalama birincil enerji tüketimi 4.661.21 exajoule olarak belirlenmiştir. CO2 emisyonundaki değişimin 98.74nün birincil enerji tüketimi ile açıklandığı ve değişen varyans probleminin varlığında kantil regresyon analizinin, doğrusal regresyon analizinden daha güvenilir sonuçlar verdiği elde edilmiştir. Kantil regresyon modeline göre; birincil enerji tüketimindeki bir exajouleluk artışın CO2 emisyonunun τ0.50. kantilinde yaklaşık olarak 54.3 milyon ton artışa sebep olacağı belirlenmiştir. Aykırı değerlerin varlığında τ0.25 ve τ0.50 için elde edilen kantil regresyon modelleri en uygun modeller olup birbirine oldukça yakın sonuçlar vermiştir. Bu modellerin, CO2 emisyonu ve birincil enerji tüketimi arasındaki ilişkinin modellenmesinde çıkarsama amaçlı kullanılması önerilmektedir.

Article Details

Investigation of the Relationship Between CO2 Emission and Primary Energy Consumption with Quantile Regression Model

Authors:

Hayriye Esra Akyüz

Publisher:

Duzce University

Main Topics:

Mühendislik

Journal:

Düzce Üniversitesi Bilim ve Teknoloji Dergisi Vol.11 Issue 3, pp.1529-1545

Document Type:

Article

Publication Date:

7/2023

Document Language:

Turkish

Document ID (Org No.):

2001296367

pISSN:

2148-2446

Primary energy consumption

CO2 emission

Linear regression

Quantile regressionBirincil enerji tüketimi

CO2 emisyonu

Doğrusal regresyon

Kantil regresyon

Issues

Article Details

Authors: Hayriye Esra Akyüz

Publisher: Duzce University

Main Topics: Mühendislik

Journal: Düzce Üniversitesi Bilim ve Teknoloji Dergisi Vol.11 Issue 3, pp.1529-1545

Document Type: Article

Publication Date: 7/2023

Document Language: Turkish

Document ID (Org No.): 2001296367

pISSN: 2148-2446

Issues

Authors:

Hayriye Esra Akyüz

Publisher:

Duzce University

Main Topics:

Mühendislik

Journal:

Düzce Üniversitesi Bilim ve Teknoloji Dergisi Vol.11 Issue 3, pp.1529-1545

Document Type:

Article

Publication Date:

7/2023

Document Language:

Turkish

Document ID (Org No.):

2001296367

pISSN:

2148-2446