<![CDATA[An Intrusion Detection System (IDS) is a computer system that analyzes user traffic and data on a network. IDS will provide a warning signal to network users that are attacked by intruders. However, IDS still needs to be developed due to the ever-evolving attack pattern. Feature Selection and Classification Network data are important components of IDS development to identify attack patterns based on network traffic data. To identify attack patterns, many researchers have proposed Feature Selection methods such as Chi-square, Gini Index, Information Gain, Relief, Gain Ratio, and Uncertainty. Researchers have also proposed Machine Learning Classification methods such as Random Forest, SVM, Naive Bayes, AdaBoost, and Neural Network. However, among the methods they propose, the Selection and Classification of Features method, the best, has not been studied. Therefore, we propose to conduct a comparative study of this method to find a better method of selection and classification of features. This study compares Feature Selection and Classification methods to identify those suitable for IDS by comparing Recall, Precision, F1, Accuracy, and Area Under Curve (AUC) for each method. The results showed that the Gain Ratio Feature Selection method was better than other Feature Selection methods. In addition, this study shows that the AdaBoost classification method is more accurate than other classification methods.]]>
Copyrights © 2025
