MLP Enhanced CO₂ Emission Prediction Model with LWSSA Nature Inspired Optimization.

Area of Science:

• Environmental Science
• Computer Science
• Data Science

Background:

• Rising carbon dioxide (CO₂) emissions pose a significant environmental threat, demanding advanced prediction methods.
• Machine learning (ML) offers powerful tools for analyzing complex emission patterns and identifying influential factors.
• Traditional optimization methods for ML models often suffer from limitations like premature convergence.

Purpose of the Study:

• To develop and validate a novel hybrid ML framework for precise CO₂ emission prediction.
• To enhance the optimization process for ML models, overcoming limitations of existing techniques.
• To identify and rank the most significant factors contributing to CO₂ emissions.

Main Methods:

• A hybrid framework combining a Multi-Layer Perceptron (MLP) with an enhanced Locally Weighted Salp Swarm Algorithm (LWSSA).
• The LWSSA integrates a Locally Weighted Mechanism (LWM) and a Mutation Mechanism (MM) for improved optimization.
• Permutation feature significance analysis was employed to determine factor importance.

Main Results:

• The LWSSA-MLP framework achieved a high prediction accuracy of 97%.
• The proposed model demonstrated superior performance compared to traditional optimizer-based MLP models.
• Key drivers identified include global trade, coal energy, export levels, urbanization, and natural resources.

Conclusions:

• The study presents a reliable and scalable framework for CO₂ emission prediction.
• The findings provide valuable insights for developing targeted interventions and sustainable development strategies.
• The enhanced LWSSA-MLP model offers a robust solution for environmental monitoring and policy-making.