EXPLAINABLE AI-BASED ANOMALY DETECTION FOR CYBERSECURITY IN CLOUD COMPUTING ENVIRONMENTS

Abstract

An increase in dependence on technology and internet based service providers has led to growth of cloud computing infrastructure across multiple sectors. Cloud-based systems have grown in popularity as organizations migrate their applications and data to cloud platforms because they offer scalable and flexible solutions. In addition, migration to cloud platforms has created additional risk areas for cyber attackers to exploit. A research study will propose the use of Explainable Artificial Intelligence (XAI) based anomaly detection system in order to improve cybersecurity in cloud computing. Machine Learning (ML), specifically traditional approaches, operate like "black box" models. Black box models limit ML's application in security critical systems that require interpretable information. The proposed method utilizes explainability techniques (SHAP & LIME) along with Deep Learning Models to provide interpretability. Results from experiments utilizing benchmark datasets (CICIDS2017) indicate that the proposed method can identify anomalous network behavior effectively while producing transparent decisions. These results show that using XAI improves trust, interpretability, and overall performance in cloud-based IDS systems.

Keywords : Cyber Security, Cloud Computing, Machine Learning, Phishing Detection & Prevention