SecRASP (Securing Runtime Applications with High-performance and Accurate RASP) is a security framework designed to enhance the protection of logistics networks by detecting zero-day attacks using machine learning models. The system integrates three powerful algorithms: Support Vector Machine (SVM), Random Forest, and XGBoost, to classify security threats in real-time. By focusing on the detection of previously unknown attack patterns, the project aims to create an adaptable security solution that can learn and detect new threats dynamically. A web-based interface developed using Flask allows users to interact with the system for registering, logging in, viewing performance metrics, and classifying attacks. This project uses a dataset from a logistics network, where real-world attack logs are analyzed and processed to train the models. The system offers comprehensive threat classification and real-time monitoring capabilities while maintaining high detection accuracy. By combining high-performance machine learning models with efficient runtime application self-protection, SecRASP ensures a robust layer of security that is capable of protecting the application layer without external intervention.

Keywords: SecRASP, machine learning, zero-day attacks, security framework, SVM, Random Forest, XGBoost, Flask, classification, logistics networks.