Artificial Intelligence (AI)-Enhanced Security Monitoring and Threat Detection in Cloud Infrastructures
Ramesh Bishukarma, Sumeet Mathur, Sandeep K. Gupta · 2025
Threat identity and counteraction measures are critical to data security and system integrity in the rapidly evolving environment of cloud computing. The current paper proposes a method to improve security monitoring and threat detection in cloud environments with ML algorithms. The approach involves data pre-processing on a data set obtained from KDD Cup’99 comprising of data imputation for missing values, normalisation using min-max scaling respectively, and feature selection by employing RFE for the efficient execution of the model. The data is divided into training and testing datasets and the proportion used is 80% training and 20% test. Classification is done with Isolated Heuristic Neural Networks (IHNN), as well as with GRU and Random Forest methods. As seen from the result analysis, the IHNN model established a high performance with a 99.35% accuracy, 96.70% precision, 91.20% recall, and a 96.32% F1 score. The efficiency of the model is then measured with parameters such as accuracy and error rate, precision, recall rate and F-score and confusion matrix in order to study errors committed. IHNN outperforms other comparable models, such as GRU and RF, in the context of threat detection in cloud environments, thus providing assured security analysis.