Logistics Delay Risk Prediction Model with Smart IoT Data: An Approach to Enhancing Supply Chain Resilience
Logistics Delay Risk Prediction Model with Smart IoT Data: An Approach to Enhancing Supply Chain Resilience
faezeh mokarrami pour1
1) M.A. Student in Entrepreneurship, Small Business Concentration, Islamic Azad University, Electronic Branch.
IOI Code
:
ABDE-ABKGCC
Publication :
6th International Conference on Advanced Research in Management And Humanities(huconf.com)
Abstract :
Delivery delay is one of the most common and costly operational disruptions in supply chains, which can propagate to other parts of the supply chain through increased transportation costs, disruption of inventory planning, and reduced service levels. Reactive approaches that identify delays after they occur provide limited opportunity for corrective intervention; hence, early prediction of delay risk is a prerequisite for operationalizing supply chain resilience.
The aim of this research is to present and evaluate a machine learning model for predicting logistics delay risk as a binary variable, using a combination of environmental-sensor data (temperature, humidity, traffic status) and operational data (inventory level, waiting time, asset utilization, and demand forecast). From an initial set of sixteen variables in a public dataset (1,000 records), seven variables were selected based on the criterion of availability prior to delay occurrence in order to control the risk of data leakage; four classification algorithms—logistic regression, decision tree, random forest, and XGBoost—were compared using a stratified train/test split (800/200).
The results showed that all four models had acceptable performance in distinguishing delayed samples (ROC-AUC in the range of 0.76 to 0.79); logistic regression showed limited relative superiority, and heavy traffic status was identified as the strongest predictive factor in all four models. The findings suggest that the output of such a model can provide a practical basis for designing a multi-level early warning system, although the relatively limited Recall of the models and the secondary, small-scale nature of the data used highlight the need for validation on real-world, larger-scale datasets prior to any operational deployment.
Keywords : logistics delay risk prediction internet of things machine learning supply chain resilience early warning system