A Secure Blockchain Framework for Privacy Preserving in Supply Chain Management based on Hybrid Autoencoder and LSTM Neural Network

Abstract

Risk identification and mitigation are essential for maintaining resilience and efficiency in the ever-changing field of supply chain management (SCM). The intricacies and uncertainties inherent in modern supply networks are often too complex for traditional risk management techniques to effectively address. In order to enhance risk detection and management in supply chain management, this study explores a hybrid strategy that blends blockchain technology with deep learning. Blockchain ensures data integrity and transparency by offering a transparent and decentralized system for supply chain operations monitoring. This process is improved by deep learning, which analyses vast volumes of historical and current data to identify patterns, predict threats, and suggest countermeasures. The proposed system leverages the inviolability of blockchain technology and the predictive capabilities of deep learning to address important challenges such as fraud detection, demand forecasting, supplier evaluation, and disruption prediction. The dataset is secured with the use of hybrid Autoencoder and LSTM based deep neural network. Autoencoder is used for dimensionality reduction and reducing the noisy as well as redundant data which is further passed through LSTM based neural network to enhance the security over the blockchain based transaction data.