A Blockchain-Enabled Vaccine Supply Chain Model Driven by Transfer Learning-Based Sentiment Analysis and Demand Forecasting

In the aftermath of global health emergencies like the COVID-19 pandemic, timely and proportionate vaccine delivery depends on efficient management of the vaccine supply chain (VSC). Conventional VSC systems often struggle to adapt to public sentiment and demand. We propose a novel architecture that optimizes vaccination delivery by combining blockchain technology, demand forecasting, and sentiment analysis to address the existing gap. For sentiment analysis, we trained a BERT-based natural language processing (NLP) algorithm incorporating transfer learning on the UCI Drug Review dataset. The model obtained 98% accuracy and an F1 score of 0.96, enhancing accuracy by 10% while requiring 50% less training data than previous approaches. In succession, a hybrid CNN-LSTM model with transfer learning was used to COVID-19 vaccination data for demand forecasting, yielding 99% prediction accuracy, a 12% improvement over baseline models, and a 36% reduction in computational time. The above models were integrated within a blockchain-based VSC system that used sentiment analysis and demand forecasting analytics to ensure safe, readily available, and efficient vaccine distribution. The proposed approach makes significant contributions in three key areas: (1) precise, data-efficient sentiment analysis that supports supply decisions, (2) improved demand forecasting with lower computational overhead, and (3) a blockchain-enabled VSC model that incorporates forecasting knowledge for safe and adaptive vaccine distribution. This study proposes a scalable and adaptive approach to prospective public health logistics, emphasizing the possibility of incorporating sophisticated AI techniques alongside blockchain for improved vaccine management.