Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{197601,
        author = {PRASANTH T and RUBEENADHARSHINI T and RAMESHWARI V and REKA S and KEERTHANA G},
        title = {Anomaly Detection in Network Traffic Using Advanced Deep Learning Techniques},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {7415-7422},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=197601},
        abstract = {In these difficult times of dealing with many interrelated systems and more devices than traditional network connections being added to the internet, accurate detection of anomalies is challenge facing security experts. Classical and rule-based IDS have numerous weaknesses, including their high false positive rates, difficulty in scaling as new devices are added and limited ability to adapt as cyber threats continue to evolve. In this article, we present a new set of techniques for implementing a deep-learning-based IDS solution for the intelligent detection of anomalies in network traffic. The framework consists of Time-Aware Normalization for normalizing Anatolated features of network traffic, Correlation-Based Feature Selection (CFS) that selects the most useful non-redundant attributes for training and uses CFS derived from multiple attributes to help minimize computational overhead when training a model, and a Long Short-Term Memory (LSTM) based CLS model that automatically detects both sequential and temporal relationships between successive network flows in order to accurately detect previously identified attacks and also detect newly developed types of attacks against computer systems. Our experiments show that our technique outperforms conventional methods with regards to accuracy, precision, recall, and F1-score across multiple benchmark datasets in an independent evaluation setting. Our framework also provides the ability to support the real-time operationalization and integration into existing infrastructure, thus providing a more comprehensive, reliable, and adaptive solution for enhancing proactive network security through the identification of normal variations and distinguishing them from malicious behaviors, which results in fewer false positives generated by the initiative.},
        keywords = {Network Anomaly Detection, Intrusion Detection System (IDS), Deep Learning, LSTM, Time-Aware Normalization, Feature Selection, Cybersecurity, Real-Time Detection},
        month = {April},
        }