The International Journal in Economics and Business Strategies
Volume 5, Numéro 2, Pages 30-38
2025-12-31
Ridge Regression For Analysing The Determinants Of Co₂ Emissions In Algeria Under multicollinearity (1970–2018)
Authors : souhila lazali . Fatiha Mezoued .
Abstract
Abstract: This study applies Ridge regression to analyse the economic and demographic determinants of CO₂ emissions in Algeria over the period 1970–2018, in the presence of severe multicollinearity that undermines the properties of the Ordinary Least Squares (OLS) estimator. The explanatory variables include GDP per capita, energy use per capita, foreign direct investment, total population, and urbanisation rate, within a log-log specification. Multicollinearity diagnostics reveal VIF values exceeding 160 for most regressors and a condition number of 37.74, rendering OLS estimates highly unstable. The shrinkage parameter k is selected via ten-fold cross-validation; Ridge regression achieves CV-MSE = 0.0206, representing a 19.1% improvement over OLS. Energy use per capita emerges as the sole statistically significant determinant (elasticity = 0.479; p < 0.01), reflecting Algeria's heavy dependence on fossil fuels.
Keywords
Keywords: Ridge regression, multicollinearity, CO₂ emissions, Algeria, bias-variance trade-off, cross-validation.
Les articles similaires
Benelbar M'hamed
.
ص 149-167.
Hacen Kahoui
.
Abdelkeder Sahed
.
pages 589-606.
Slimani Salim
.
pages 73-89.
Bouhali Ihssane
.
Djemaci Brahim
.
pages 378-394.
Boufenneche Wassila
.
pages 22-36.
