ENHANCING SUPPLY CHAIN AGILITY AND SUSTAINABILITY THROUGH MACHINE LEARNING

Consequently, the present research study aims to explore how machine learning prospects enhance the dynamic and resilience of the supply chain networks that have been affected by modern circumstances. In this case, both forms of machine learning; supervised and unsupervised were applied; linear regression, recurrent neural networks LSTMs, and feedforward neural networks with attention. Thus, this paper also describes how these models can be useful in improving demand forecasting, risk assessment, and inventory management. We discovered that extended linear regression was well appropriate for short-time demand forecasting the moment limitations became visible when the model had to handle intricate and multiple tasks. In contrast to the other models, the LSTM models are much higher in terms of time series forecasts on the fluctuations and helped minimize the shock of the supply chain. An enhanced feedforward neural network with attention mechanisms enhanced the interpretative ability of the model and the important aspects that are required in transporter delays or suppliers’ performances. Data preprocessing was okay to preprocess being to have the models run at their optimum. Altogether, the present research provides a comparative analysis of different approaches to the classification of data using the methods of machine learning that are essential for the improvement of the models of supply chain that seem to be more resistant and efficient. Thus, supplementing the findings of existing studies that examined the use of predictive analytics in supply chain management, the ideas put forward in this paper provide evidence for the opinions that can be found in previous work regarding the growing importance of machine learning to address new challenges in supply chain management.