Analysis of intrusion detection system in cloud computing environment using artificial neural network

Abstract

This study developed a novel intrusion detection system (IDS) for cloud computing using artificial neural networks (ANNs) and machine learning techniques. The proposed IDS uses an adaptive architecture capable of detecting malicious activities within a cloud computing environment. To process and optimize the data, Adam optimization techniques were employed, and MiniMaxScaler was used to normalize the data for training. The model was designed using the TensorFlow framework for ANNs, and the LSD methodology was employed in the development. The training was conducted using the University of New Brunswick Intrusion Detection Systems dataset, which had been preprocessed. Results indicate that the proposed architecture was highly effective in detecting various attacks, with low false-positive and false-negative rates. The training and validation accuracies were 99.7% and 99.9%, respectively, using this method. This approach can automatically detect the nature of attacks, saving time and resources.