Browsing by Author "Michael Favour Edafeajiroke"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    An Enhanced Network Intrusion Detection System using Dimensionality Reduction Techniques
    (KASU Journal of Computer Science, 2024) Olufemi James Akinola; Sulaiman Olaniyi Abdulsalam; Michael Favour Edafeajiroke; Kazeem Alagbe Gbolagade
    Protecting the privacy and confidentiality of data in computer networks is crucial, necessitating reliable intrusion detection methods. The complexity and vast volume of data present challenges to effective intrusion detection, as traditional models struggle with high-dimensional data that reduce accuracy. This study addresses these challenges by integrating Mutual Information (MI) and Linear Discriminant Analysis (LDA) as feature selection and extraction techniques respectively, to enhance the classification performance of Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) classifiers. The Canadian Institute for Cybersecurity Intrusion Detection System 2017 (CIC-IDS 2017) dataset was used to develop an Intrusion Detection System (IDS). The Results findings revealed that integrating MI+LDA with SVM yielded an accuracy of 98.84% while combining MI+LDA with KNN produced an accuracy of 90.17%. The result of the MI+LDA+SVM model based on accuracy, sensitivity, specificity, and F1-score matrixes, outperformed that of MI+LDA with KNN and other existing models. Therefore, the incorporation of MI and LDA dimensionality reduction techniques proved reliable for boosting the performance of an intrusion detection system in computer networks. Future research could apply the result of this study to other types of computer network benchmark datasets and other kinds of machine learning techniques.
  • Thumbnail Image
    Item
    Development of an Intrusion Detection System using Mayfly Feature Selection and Artificial Neural Network Algorithms
    (LAUTECH Journal of Engineering and Technology, 2024-06-24) Sulaiman Olaniyi Abdulsalam; Rasheed Abiodun Ayofe; Michael Favour Edafeajiroke; Jumoke Falilat Ajao; Ronke Seyi Babatunde
    Protecting the privacy and confidentiality of information and devices in computer networks requires reliable methods of detecting intrusion. However, effective intrusion detection is made more difficult by the enormous dimensions of data available in computer networks. To boost intrusion detection classification performance in computer networks, this study proposed a feature selection mode for the classification task. The proposed model utilized the Mayfly feature selection algorithm and ANN as the classifier. The model was also tested without a Mayfly Algorithm (MA). The efficiency of the model was determined through a comparison of its Accuracy, Specificity, Precision, Negative predictive value, False positive rate, False discovery rate, False negative rate, Sensitivity, and F1- score. The experimental outcomes revealed that the proposed model is more efficient than existing models when implemented on the Canadian Institute for Cybersecurity Intrusion Detection System 2017 (CIC-IDS 2017) dataset. Accuracy of 99.94% (using Data+mayfly+ANN) and 90.17% (using Data+ANN) were attained after experimentation. The proposed model demonstrated superior accuracy compared to existing studies. Its robustness is due to employing mayfly techniques that combine the strengths of PSO, GA, and FA for optimal feature selection. This research presents a dependable dimensionality reduction model useful for intrusion detection and improving security in computer networks.