An Efficient Network Intrusion Detection System Using ANN and SVM with Feature Selection

Authors

  • Y.Nagamalleswararao 1 ,K.Pavani2 ,S.Dharani Sri Lakshmi Bashitha3 Author

DOI:

https://doi.org/10.62643/

Abstract

Due to cyberattacks, network security has become a major issue as internet-based services proliferate. This research introduces an efficient Network Intrusion Detection System (IDS) that uses supervised machine learning and feature selection to categorize network traffic as normal or malicious. The suggested system uses SVM and ANN algorithms to choose the most relevant features, reduce dimensionality, and enhance model performance utilizing filter and wrapper techniques. The labeled NSL-KDD dataset trains and evaluates the system on network assaults and regular traffic. Before model training, data is cleaned, encoded, and normalized. Experimental findings show that the ANN model with wrapper-based feature selection outperforms the SVM model in detection and classification. Improved accuracy and fewer false positives make the proposed intrusion detection method more useful for real-world network security applications. Detecting unknown or zero-day assaults is difficult, requiring significant deep learning research.

Downloads

Published

30-05-2026

How to Cite

An Efficient Network Intrusion Detection System Using ANN and SVM with Feature Selection. (2026). International Journal of Engineering Research and Science & Technology, 22(2(2), 738-746. https://doi.org/10.62643/