A Decision-Making System for Medical Transportation Mode Using Machine Learning Methods

The freight transportation system faces complex operations globally to meet customer demands. Intense competition prompts companies to enhance performance. Transportation modes (road, sea, air) impact service levels, each with distinct features, benefits, costs, environmental effects, and societal risks. Shippers confront challenges in mode selection due to numerous factors, compounded by an increase in low-volume, high-frequency shipments. Rising logistics costs for a few products of exporters affect the socio-economic situation of a country. This research introduces a hybrid approach for a shipment selection model, focusing on pharmaceutical drugs. Utilizing ma- chine learning algorithms (decision tree, Random Forest, logistic regression, XGboost, SVM) and multi-criteria decision-making methods (SAW, MARCOS, TOPSIS, MULTIMOORA, VIKOR), this study predicts the optimal shipping method (land, air, sea) based on dataset features (shipping cost, origin-destination, cargo weight, dimensions). Evaluation metrics include F1 score, Recall, Precision, and Accuracy score. XGboost stands out as the optimal algorithm, demonstrating an accuracy of eighty-four percent, with random forest, decision tree, SVM, and logistic regression following in descending order. This comprehensive approach addresses the complexities of pharmaceutical shipment selection, considering various influential factors.