A Hybrid Learning Model for Detecting Attacks in Cloud Computing

Cloud computing has revolutionized how organizations manage and deliver IT services, offering unprecedented scalability and flexibility. However, with this transformation, the cloud infrastructure has become a prime target for cyber-attacks. Detecting and mitigating attacks in cloud computing environments is paramount to ensure data security and business continuity. This paper presents a novel Hybrid Learning Model (HLM) for detecting attacks in cloud computing. The proposed HLM leverages the strengths of both supervised and unsupervised learning techniques to enhance attack detection accuracy and reduce false positives. The proposed approach is the combination of Iterative Intrusion Detection system (IIDS) combined with random forest (RF) as the training model and iterative reinforcement learning based Support Vector Machine (SVM) with transfer Learning is used to classify the types of attacks. The preprocessing techniques used in this paper are Modified Z-Score and SMOTE (Synthetic Minority Over-sampling Technique) used to clear the data. A feature selection technique Correlation-based Feature Selection (CFS) is also used to select the accurate features from the data. It combines the discriminative power of supervised machine learning with the anomaly detection capabilities of unsupervised learning. The model is trained on a diverse dataset of regular and attack traffic to learn the patterns and behaviors of various attacks in a cloud environment. The HLM is evaluated using real-world datasets from cloud environments and compared to traditional detection methods. Our results show that the Hybrid Learning Model significantly improves attack detection accuracy while reducing false positives. This research contributes to the ongoing efforts to enhance the security of cloud computing infrastructures and protect critical data assets.