Aug 2026· FUDMA Journal of Sciences· 0 citations· 1 references
TL;DR
The findings indicate that hybrid deep learning techniques can improve network security by enhancing intrusion detection capability while reducing false alarms.
Abstract
The rapid growth of digital communication technologies and interconnected network infrastructures has increased the frequency and sophistication of cyber threats. Traditional Network Intrusion Detection Systems (NIDS), which primarily depend on signature-based and rule-based approaches, often struggle to detect zero-day attacks and evolving intrusion patterns. This study presents the design and implementation of a hybrid Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) deep learning-based Network Intrusion Detection System to improve the detection accuracy and reliability of malicious network activities. The proposed system combines the feature-extraction capabilities of CNNs with the temporal sequence-learning capabilities of LSTMs to identify both spatial and sequential characteristics of network traffic. The CICIDS2017 benchmark dataset was utilised for system training and evaluation. Data preprocessing techniques such as removing missing values, feature scaling, label encoding, and Synthetic Minority Oversampling Technique (SMOTE) balancing were applied to improve data quality and class distribution. The model was implemented using Python, TensorFlow, Keras, Scikit-learn, NumPy, and Pandas within a Google Colab environment. The processed dataset was partitioned into 80% for training and 20% for testing. Experimental results demonstrated high intrusion detection performance with an accuracy of 98.48%, precision of 97.75%, recall of 99.23%, F1-score of 98.49%, and a false alarm rate of 2.28%. In addition to the deep learning model, a web-based user interface was developed to support traffic prediction, performance monitoring, and management of prediction history. The findings indicate that hybrid deep learning techniques can improve network security by enhancing intrusion detection capability while reducing false alarms.
The rapid growth of digital communication technologies, cloud computing, and Internet of Things (IoT) devices has
increased both the frequency and sophistication of cyber-attacks, making effective intrusion detection an essential component of
modern cybersecurity systems. Traditional signature-based intrusion detection systems (IDS) are effective against known attacks
but fail to detect previously unseen or evolving threats. This study investigates the application of deep learning models for binary
network intrusion detection using the NSL-KDD benchmark dataset. Three standalone architectures, Convolutional Neural
Networks (CNN), Long Short-Term Memory (LSTM) networks, and Deep Neural Networks (DNN), are implemented and
evaluated, alongside a CNN-LSTM Hybrid model that integrates spatial and sequential learning, and a DNN-LSTM Ensemble
model that combines independently trained DNN and LSTM predictions through weighted averaging. Following data cleaning,
categorical encoding, normalization, and Random Forest-based feature selection (41 features reduced to 20), all models were
trained and evaluated under identical conditions using Accuracy, Precision, Recall, F1-Score, ROC-AUC, training time, and
inference time. The standalone DNN model achieved the best overall performance, with 80.98% accuracy, 97.08% precision,
68.66% recall, 80.43% F1-score, and 96.11% ROC-AUC, while also requiring the shortest training time (39.69 s). The CNNLSTM Hybrid model attained the highest precision (97.23%) but did not outperform the standalone architectures overall, and
the DNN-LSTM Ensemble produced balanced but not superior results. These findings indicate that carefully designed
standalone architectures can match or exceed the performance of more complex hybrid and ensemble models for binary
intrusion detection, while incurring substantially lower computational cost. The study contributes a controlled, commonframework comparison of five deep learning architectures and provides practical guidance for selecting computationally
efficient models for anomaly-based intrusion detection.
Ketki Naik, Sanjeev Ghosh· International Journal for Re...· 0 citations
This study examines a one-dimensional Convolutional Neural Network and a hybrid model, investigating how both architectures can detect network attacks in binary and multiclass classification settings, and provides actionable insights for practitioners choosing between deep learning and classical approaches under real-world NIDS deployment constraints.
The interconnected systems, cloud infrastructures, Internet of Things (IoT) devices, and large networks of today have increased the complexity and frequency of cyberattacks. There is now a growing requirement for intelligent and adaptive Intrusion Detection Systems (IDS). Classic signature-based and machine learning-based IDS methodologies usually fail to recognise advanced, evolving, or zero-day attacks due to their single-feature extraction capabilities and reliance on handcrafted rules. Accordingly, Deep Learning (DL) methods have surfaced with the ability of learning features automatically for better detection performance in other networks. This survey presents recent developments that have taken place in deep learning-based intrusion detection systems with supporting statistics from thirty-six representative works published across IoT, cloud, enterprise and SDN. This paper systematically develops a taxonomy of IDS architectures and a survey of recently proposed networks, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), LSTMs, GRUs, Autoencoders, GANs, Transformer-based, as well as hybrids. The performance of models has been evaluated using popular benchmark datasets such as NSL-KDD, UNSW-NB15, CICIDS2017, CICIDS2018, N-BaIoT and CICIoT2023. Many studies reported a detection accuracy more than 98%. Innovative methods such as feature selection, data augmentation, hyperparameter optimisation and spatial temporal learning have been discussed in detail. Although strides have been made yet numerous challenges still loom large, from class imbalance to high computational costs and limited interpretability of models. In the conclusion section, we highlight some of the important research challenges and future directions. Further, we discuss explainable artificial intelligence, adversarial robustness, federated learning, lightweight edge-deployable models, and Transformer-based IDS frameworks. According to this survey, this review will help researchers and practitioners in understanding the current developments and in designing the next generations intelligent intrusion detection systems.
Sajith K. V., Gripsy Paul, Bhagavant Deshpande et al.· Journal of Network and Infor...· 0 citations
A deep learning-based intrusion detection framework that utilizes flow-level statistical features and a hybrid Convolutional Neural Network–Bidirectional Long Short-Term Memory (CNN–BiLSTM) architecture to capture both spatial and temporal characteristics of network traffic is proposed.
H. Hussein, Saad Raad, Ali Qasim· Al-Noor Journal of Engineeri...· 0 citations
In rapid development of computer networks, anomaly detection is essential to securing modern communication networks against emerging cyber-attacks, zero-day vulnerabilities, and massive malicious traffic. But conventional intrusion detection systems are plagued by high false alarms, scalability issues and low adaptability to evolving network conditions. To overcome these challenges, this research suggests an adaptive deep learning and ensemble model for scalable real-time detection of network anomalies. This approach combines CNNs for spatial feature learning, RNNs for temporal pattern recognition, and dynamically weighted ensemble learning for making reliable decisions. The model uses normalization, feature selection and class-balancing to enhance generalization. The model is tested across benchmark datasets (CICIDS2017, NSL-KDD, UNSW-NB15) under various attack scenarios. The proposed framework achieves 97.83% accuracy, 97.8% precision, 97.6% recall, 97.7% F1-score and surpasses traditional machine learning and pure deep learning approaches. Further, the model exhibits low inference times and better adaptability to concept drift. This system can be deployed in enterprise, IoT, and edge-security systems for efficient, scalable, and intelligent threat detection.
Kavali Divya Sree, Y. Vishnu Tej· International Conference Com...· 0 citations
This article proposes an advanced method for network intrusion detection using a combination of recurrent neural networks (RNNs), specifically long short-term memory (LSTM), gated recurrent units (GRU), and bidirectional long short-term memory (BiLSTM) models, enhanced by synthetic minority oversampling technique (SMOTE) to address class imbalance in datasets like network security laboratory–knowledge discovery in databases (NSL-KDD). The method aims to accurately classify network traffic by learning temporal patterns of both normal and malicious activities. SMOTE is employed to balance the dataset, ensuring that underrepresented attack types receive adequate model attention, thereby improving model robustness. The proposed models (LSTM, GRU, and BiLSTM) are trained and evaluated on the NSL-KDD dataset, with hyperparameter tuning performed through RandomizedSearchCV for optimal performance. The results show a significant improvement in accuracy, precision, recall, and F1-score, with BiLSTM demonstrating the highest performance, achieving near-perfect classification results (99.5% accuracy). This method not only mitigates the issue of class imbalance but also leverages the power of RNNs for sequence modeling, offering a promising solution for effective intrusion detection in modern networks.
Prajwalasimha Sindugatta Nagaraja, Navya Rajashekara, Pushpa Bangalore Ramesh et al.· IAES International Journal o...· 0 citations
