Machine Learning Model for Detecting Attack in Service Supply Chain

Supply chain attacks exploit weaknesses in third-party vendors, software updates, and service providers, mainly posing a cybersecurity problem. Traditional detection methods often lag behind these sophisticated attacks. The study employs machine learning methods to increase the detection of service supply chain attacks, including Decision Trees, Random Forest, and XGBoost algorithms. These models were assessed in accordance with accuracy, precision, recall, and the F1-score, with Random Forest topping the list with an accuracy of 96.1%, followed by Decision Trees with 95.0% accuracy and XGBoost with 94.7% accuracy. Through the use of graphs showing the ROC and Precision-Recall curves, Random Forest can best describe the balance between precision and recall. Random Forest is tremendously good for detection with less false positives; however, due to its high computational costs, it may be challenging to implement in real-time. These results shed light on the potential of machine learning technology to outperform traditional intrusion detection systems and enhance cybersecurity in service supply chains. Future research will focus on real-time implementation and hybrid models that combine classical and deep learning techniques.