AI-Based Learning Model for Sociocybernetic Systems in Web of Things : An Efficient and Accurate Decision-Making Procedure

Singh, Priti and Rathee, Geetanjali and Kerrache, Chaker Abdelaziz and Bilal, Muhammad and Calafate, Carlos T. and Wang, Huihui (2024) AI-Based Learning Model for Sociocybernetic Systems in Web of Things : An Efficient and Accurate Decision-Making Procedure. IEEE Systems, Man, and Cybernetics Magazine, 10 (4). pp. 40-48. ISSN 2380-1298

Text (AI-based_Learning_Model_for_Socio-Cybernetics)
AI-based_Learning_Model_for_Socio-Cybernetics.pdf - Accepted Version
Available under License Creative Commons Attribution.
Download (426kB)

Abstract

Cybernetic threats have become a growing concern in recent years, highlighting the need for effective intrusion detection systems (IDSs) to detect and prevent social cyberattacks. Sociocybernetics is a significant platform for providing real-time mapping or to enable information access across heterogeneous networks. However, ontology-based knowledge and web support for social cybernetics demand massive warehouses that provide the required computational power for log applications and data-processing mechanisms, in addition to effective decision-support solutions for business by extracting useful information in a very secure and intelligent way. In this work, we propose an IDS approach that combines a tree-based XGBoost algorithm and a bidirectional long short-term memory (BiLSTM) network to address the limitations of traditional approaches. The proposed approach includes multiple steps, such as data preprocessing, feature selection using an infinite feature selection (IFS) algorithm, and the application of principal component analysis (PCA) for dimensionality reduction. Furthermore, a direct trust-based scheme is used to strengthen the decision-making process by improving the overall accuracy in the network. The performance of the proposed approach is evaluated based on accuracy, precision, recall, and F1 score and is compared with the existing LSTM-based deep learning model (LBDMIDS) method. Experimental results demonstrate that the proposed approach outperforms traditional methods by providing higher accuracy along with a slight improvement in terms of precision, recall, and F1 score. In particular, the proposed mechanism shows a 99% improvement in terms of accuracy compared to existing schemes, while also ensuring secure communication in the network.