Automated Cybersecurity Risk Assessment for SMEs Using Machine Learning and Public Threat Intelligence

Abstract

In the modern digital landscape, small and medium-sized enterprises (SMEs) face increasing cybersecurity risks due to limited financial resources, lack of dedicated security teams, and insufficient visibility into their threat exposure. This paper presents an intelligent and automated Cybersecurity Risk Scoring System designed to quantitatively assess the cybersecurity posture of SMEs based on their publicly accessible digital footprint. The proposed system integrates multiple threat intelligence sources, including VirusTotal, Shodan, Have I Been Pwned, and AbuseIPDB, through a unified backend API developed using Python-based Flask and FastAPI frameworks. Security-related features such as exposed network services, malware indicators, IP reputation, domain characteristics, and breach history are aggregated and analyzed using an XGBoost-based machine learning model to generate a normalized and interpretable risk score. A cross-platform Flutter-based mobile interface enables organizations to visualize domain health, vulnerability exposure, and network anomalies in real time. By automating data collection, analysis, and risk visualization, the proposed approach supports proactive cybersecurity risk management for SMEs while remaining cost-effective and scalable. The system aligns with established cybersecurity best practices and demonstrates the effectiveness of machine learning-driven risk assessment using publicly available threat intelligence data.