Predicting CO₂ emissions in Morocco: exploring the use of ridge regression with data preprocessing and feature impact...

Area of Science:

• Environmental Science
• Econometrics
• Climate Modeling

Background:

• Morocco faces challenges in managing CO2 emissions influenced by energy, economic, and demographic factors.
• Accurate prediction of carbon dioxide (CO2) emissions is crucial for effective climate change mitigation strategies.

Purpose of the Study:

• To apply and evaluate ridge regression (RR) for predicting CO2 emissions in Morocco.
• To compare the performance of RR against LASSO, ElasticNet, and linear regression models.
• To identify key drivers of CO2 emissions and assess the impact of renewable energy policies.

Main Methods:

• Utilized time-series data from 1965-2021 for Morocco, incorporating ten variables.
• Applied ridge regression (RR), LASSO regression, ElasticNet regression, and linear regression.
• Performed feature importance analysis and policy scenario testing for renewable energy targets.

Main Results:

• The RR model achieved the highest prediction accuracy (R² = 0.994), outperforming other methods.
• Oil consumption was identified as the primary driver of CO2 emissions (129% standardized effect), followed by coal (9%).
• Renewable and hydro energy generation demonstrated a negative correlation with emissions, indicating a mitigating effect.

Conclusions:

• Ridge regression is a reliable tool for environmental modeling and CO2 emission prediction.
• Achieving Morocco's 2030 renewable energy target could reduce annual emissions by approximately 0.67 Mt.
• Expanding renewable energy deployment is strategically vital for Morocco's climate mitigation efforts.