AHSWN Home • Issue Contents • Forthcoming Papers
Securing Cloud from Attacks: Machine Learning based Intrusion Detection in Cloud Sensor Networks
Meble Varghese and M. Victor Jose
“Intrusion Detection System (IDS) is the most commonly used mechanism to detect attacks on cloud sensor networks. An IDS is a type of security system designed to automatically alert administrators when someone or something is trying to compromise information system through malicious activities or through security policy violations in cloud”. This paper aims to establish a new IDS model, based on Machine Learning (ML) technology in cloud networks. It consists of two phases, namely feature extraction and classification. During the feature extraction phase, the central tendency features, as well as proposed ordinal features, are extracted. The lower, as well as higher order statistical features are then subjected to classification process, which is processed using the “Deep Convolutional Neural Network (DCNN)” framework. The major contribution is to optimally tune the weight of DCNN using a hybrid algorithm that ensures a better detection rate. For optimization purposes, this work deploys the hybridized concepts of both the “Monarch Butterfly Optimization (MBO) and Moth Search Algorithm (MSA) algorithm” together termed as BM-MSA. Finally, the proposed method provided better outcomes when compared with other traditional methods.
Keywords: Cloud Security; Ordinal features; Monarch butterfly Optimization;Moth Search Algorithm; DCNN