A Trust Aware Behavioral Based Intrusion Detection in Cloud Environment Using Ensemble Service Centric Featured Neural Network

• Published in: 2021 4th International Conference on Computing and Communications Technologies (ICCCT) pp. 342 - 349
• Main Authors: Ahmed, N.Zafer, Durga, R.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 16.12.2021
• Subjects: Behavioral analysis; Cloud computing; Communications technology; Data analysis; Feature extraction; Feature selection and classification; IDS; Intrusion detection; Recurrent neural networks; Service monitoring; Time complexity
• DOI: 10.1109/ICCCT53315.2021.9711827

The modern environment development comes under internet of service access for various purpose for doing communication. The cloud centric services are distributed to the user by access via the network. By the nature of communication be affected by various intrusion by accessing the services during wrongly manner by the intruders. The Advance Intrusion Detection System (AIDS) make effective monitoring by accessing the user service log to find the intrusion, but the prediction was not produce accuracy because of dimensionality features affect the identification of user behavior. To resolve this problem, we propose a Trust Aware Behavioral based Intrusion detection System (TABIS) to predict the user behavioral features related to malicious activity during the service of access. The system monitors the activity of the User Service Access Rate (USAR) to estimate the Trust Factor Rate (TFR). The features are selected based on mutual activity using Ensemble Service Centric Feature Selection (ESCFS) by accessing the service logs to choose the relative features. Based on the estimated trust factor weighting the features are selected and trained into Sigmoid Recurrent Neural Network (SRNN) to classifying the risk of evaluation in intrusion by class by category. The proposed system produce high intrusion detection rate as well produce best precision, recall, classification accuracy than any other methods.