MACHINE LEARNING FOR SUPPLY CHAIN RISK ANALYSIS AND BUSINESS RESILIENCE UNDER UNCERTAIN DEMAND

Objective: The study is focused on finding the relationships between demand uncertainty, supply chain-related factors, and total supply chain cost in order to improve the understanding of the risk exposure and the risk resilience-related factors. Method: In this research, risk patterns and business resilience in a pharmaceutical supply chain that operates under uncertain demand are explored through an exploratory analytical approach with the aid of machine learning. The research is based on a secondary dataset generated from a blockchain-enabled pharmaceutical supply chain (Pharma SC) environment consisting of 5000 generated observations. The distributions, differences, and inter-relationships of variables are explored using descriptive statistics, visualization methods, and correlation-based methods. Results: The results indicate that there is a significant range in the total supply chain cost, with a higher share of observations in the medium cost range and a lower share of high-cost scenarios. Comparative analysis shows that the installation of the blockchain and the variables related to the finished product show significant differences in the various cost categories, while the other variables suggest fewer differences. The results of the correlation analysis also reveal that the installation of blockchain and the finished products were found to have a higher correlation with the total supply chain cost when compared to demand uncertainty only. Novelty: The results indicated that supply chain risk is not a single source of uncertainty but rather is a result of the interaction of various operational and resource-related uncertainties. The findings underscore the significance of machine learning-driven exploratory data analysis in identifying meaningful patterns in intricate supply chain data.