부산대학교 도서관

In the current interconnected and susceptible digital environment, the demand for reliable network attack prediction systems has reached a critical level. This research paper presents a novel approach for network intrusion detection utilizing the Improved Import Vector Machine (IVM) classification algorithm combined with feature selection using Cuckoo Search based Grey Wolf Optimization (CS-GWO). The aim is to improve the accuracy and effectiveness of intrusion detection systems by selecting the most relevant features and employing a powerful classification algorithm. The proposed methodology begins with data preprocessing, ensuring the dataset is appropriately cleaned and normalized. Subsequently, the CS-GWO algorithm is applied for feature selection, leveraging the strengths of both Cuckoo Search and Grey Wolf Optimization. This hybrid approach efficiently explores the search space to identify the optimal feature subset that contributes to accurate intrusion detection. Once the feature selection process is completed, the IVM classification algorithm is employed to classify network instances into intrusion or normal categories. The IVM algorithm is chosen for its ability to handle high-dimensional data and its robustness in dealing with imbalanced datasets commonly encountered in network intrusion detection. To evaluate the performance of the proposed approach, extensive experiments are conducted using benchmark datasets. Various evaluation metrics such as accuracy, precision, recall, and F1 score are utilized to assess the effectiveness of the system in detecting network intrusions. The results demonstrate that the combination of IVM classification and CS-GWO feature selection significantly improves the accuracy and efficiency of network intrusion detection. The selected features obtained through CS-GWO contribute to enhanced classification performance, enabling the system to accurately identify and respond to potential intrusions in real-time scenarios.