Advanced Intrusion Detection in Social Networks Through Soft Computing Approaches
International Journal of Information Technology & Computer Sciences Perspectives · 30.06.2026 · 15 (1) · 46-53
DOI: 10.67591/itcs2026.15.1.62 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Information Technology & Computer Sciences Perspectives, online erschienen am 30.06.2026. Verfasst von Vijay Gupta und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Objectives
This study proposes an advanced intrusion detection framework for social and computer networks by integrating soft computing and machine learning techniques to improve intrusion detection accuracy while reducing false positive rates and computational complexity. The research aims to develop an efficient anomaly-based Intrusion Detection System (IDS) capable of identifying malicious network activities through intelligent feature selection and ensemble classification. Theoretical Framework: The study is grounded in the concepts of anomaly-based intrusion detection, soft computing, ensemble machine learning, feature selection, and intelligent classification. It integrates Genetic Algorithms (GA), Partial Decision Trees (PART), bootstrap aggregation (Bagging), fuzzy reasoning, clustering concepts, and rule-based learning to establish an adaptive framework for improving intrusion detection performance. Methodology: An experimental machine learning approach was employed to develop the proposed intrusion detection framework. Initially, a Genetic Algorithm was applied to the NSL-KDD dataset to select the 15 most relevant network traffic features for intrusion classification. These optimized features were then used to train a bagging ensemble based on Partial Decision Tree (PART) classifiers. The proposed model was evaluated using both 10-fold cross-validation and an independent test dataset. Performance was compared with conventional classifiers, including Naïve Bayes, PART, C4.5, Bagging (Naïve Bayes), and a hybrid C4.5+SVM model using Root Mean Square Error (RMSE), True Positive Rate, False Positive Rate, and Classification Accuracy.
Findings
Experimental results demonstrate that the proposed ensemble classifier consistently outperformed the comparison methods. During cross-validation, the proposed framework achieved a classification accuracy of 99.71%, exceeding the performance of standalone PART (99.66%) and C4.5 (99.69%) classifiers while maintaining a low false positive rate of 0.003. On the independent test dataset, the proposed model achieved an accuracy of 79.87%, outperforming the hybrid C4.5+SVM approach (75.55%) and other baseline classifiers. The integration of Genetic Algorithm-based feature selection with bagged Partial Decision Trees significantly reduced computational complexity and improved detection reliability. Research Implications: The proposed framework provides practical guidance for cybersecurity professionals and network administrators by demonstrating the effectiveness of combining feature optimization with ensemble learning to develop accurate, scalable, and computationally efficient intrusion detection systems. The findings also establish a foundation for future research integrating advanced deep learning and adaptive soft computing techniques for real-time network security. Value: This study contributes to intrusion detection research by proposing and experimentally validating a hybrid soft computing framework that combines Genetic Algorithm-based feature selection with bagging of Partial Decision Trees. The proposed approach enhances classification accuracy, minimizes false alarms, and improves computational efficiency, providing an effective solution for next-generation intelligent intrusion detection systems.
Zitieren
APA
Gupta, V., Gupta, N., Kumar, R., & Dangi, A. (2026). Advanced Intrusion Detection in Social Networks Through Soft Computing Approaches. International Journal of Information Technology & Computer Sciences Perspectives, 15(1), 46-53. https://doi.org/10.67591/itcs2026.15.1.62
Vancouver
Gupta V, Gupta N, Kumar R, Dangi A. Advanced Intrusion Detection in Social Networks Through Soft Computing Approaches. International Journal of Information Technology & Computer Sciences Perspectives. 2026;15(1):46-53. doi:10.67591/itcs2026.15.1.62
BibTeX
@article{Gupta2026,
author = {Gupta, Vijay and Gupta, Nikita and Kumar, Rajesh and Dangi, Abhilasha},
title = {Advanced Intrusion Detection in Social Networks Through Soft Computing Approaches},
journal = {International Journal of Information Technology & Computer Sciences Perspectives},
year = {2026},
volume = {15},
number = {1},
pages = {46--53},
doi = {10.67591/itcs2026.15.1.62}
}
EUVIMED