A META-ANALYSIS OF DEEP LEARNING-BASED ECONOMIC RECOVERY FRAMEWORKS FOR SUSTAINABILITY AND CLEAN ENVIRONMENT INITIATIVES USING IOT TECHNOLOGIES

This meta-analysis explores the integration of Deep Learning (DL) and Internet of Things (IoT) technologies within the context of sustainable economic recovery and clean environmental initiatives. As global economies seek resilient and low-carbon pathways in the aftermath of economic disruptions and environmental crises, the role of intelligent technologies in supporting data-driven decision-making has become increasingly critical. This study systematically reviewed 147 peer-reviewed articles published between 2010 and 2025, encompassing empirical research across sectors such as energy, transportation, agriculture, urban planning, and environmental monitoring. The findings reveal that over 70% of the reviewed studies reported significant improvements in prediction accuracy, operational efficiency, emissions reduction, and resource optimization when DL models were applied to real-time data generated by IoT infrastructures. DL architectures such as LSTM, CNNs, and transformers consistently outperformed traditional forecasting models in dynamic and multivariate settings. However, several persistent challenges were identified, including issues of dataset bias, model transparency, high energy consumption in DL training, and limited access to digital infrastructure in developing regions. The review also highlights the lack of standardized evaluation metrics and governance frameworks, which impedes scalability and cross-sector benchmarking. Despite these limitations, the overall evidence supports the transformative potential of DL-IoT systems as intelligent enablers of green recovery strategies. This study offers a comprehensive synthesis of current applications, challenges, and sectoral impacts, contributing valuable insights for researchers, practitioners, and policymakers aiming to leverage emerging technologies for sustainable development and environmental resilience.