Background and objective: Tracking clean electricity generation in developing economies is highly challenging owing to the influence of turbulent external factors. Clean electricity is a significant enabler of striving toward environmental sustainability. In this research, we aim to model hydro, nuclear, and renewable electricity generation in India through applied predictive modeling. We also strive to uncover the influence of the critical determinants responsible for clean electricity growth.
Methodology: We propose a granular predictive framework comprising ensemble empirical mode decomposition, clustering applications in spatial data based on density, including noise, and atom search optimization-based novel optimization methodology to predict absolute figures of clean energy generation. The framework uses a series of socio-economic factors reflecting household demand and industrial growth in India as explanatory variables.
Results: The rigorous scrutiny of the predictive framework specifies hydro electricity generation is relatively more predictable during the time horizon influenced by the COVID-19 pandemic. The deployment of dedicated explainable artificial intelligence (AI) tools suggests an increased adoption of clean electricity in selected industrial sectors in India, which broadly governs the evolutionary pattern.
Conclusion: The underlying research is the first of its kind to fathom the daily temporal dynamics of clean electricity generation in the Indian context. Consideration of three distinct clean electricity sources during highly volatile time regimes underscores the contribution of the work. The predictive framework survives a stringent performance check, which justifies the robustness of the same. Demand in different industrial sectors in India profoundly influences the growth toward clean electricity.
Keywords: Atom search optimization; Clean electricity; Empirical ensemble mode decomposition; Explainable artificial intelligence; Industrial growth.
© 2023 The Authors.