What question did this study set out to answer?

May 7, 2026Journal of Engineering0 citationsOpen Access

A Deep Learning Approach Using Optimized LSTM for Anomaly‐Based Network Intrusion Detection

Key Points

This research aims to enhance anomaly-based network intrusion detection systems using an optimized LSTM model.
Developed an optimized long short-term memory (LSTM) model for anomaly detection.
Incorporated synthetic minority oversampling technique (SMOTE) to address class imbalance.
Evaluated the model using the CICIDS2017 benchmark dataset.
The optimized LSTM model outperformed traditional machine learning algorithms.
Improvements observed in accuracy, precision, recall, and F1-score.
Highlights the potential of advanced LSTM architectures in intrusion detection.

Abstract

With the exponential rise in cyber threats, anomaly‐based network intrusion detection systems (NIDSs) have become critical for maintaining robust cybersecurity. This article proposes an optimized long short‐term memory (LSTM) deep learning (DL) model specifically designed to detect anomalies in network traffic. By leveraging temporal dependencies and sequential patterns, the model improves the detection of complex and evolving intrusion behaviors. To further enhance performance, the framework incorporates synthetic minority oversampling technique (SMOTE) and feature selection to address class imbalance and reduce dimensionality. Evaluation using the CICIDS2017 benchmark dataset demonstrates that the proposed model outperforms traditional machine learning (ML) algorithms in terms of accuracy, precision, recall, and F1‐score. The findings highlight the potential of advanced LSTM architectures in significantly enhancing the effectiveness of intelligent IDSs.

AI에게 질문

Bookmark

View Full Paper