Comparative Approach for Intrusion Detection using CNN

Abstract

In computer network security, intrusion detection plays an important role in protecting systems from various threats. However, as intrusion techniques become more sophisticated, traditional detection methods have shown limitations in identifying new attacks. This research addresses the pressing issue of improving intrusion detection by utilizing Convolutional Neural Networks (CNN) algorithms, compared to various other machine learning techniques such as Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Gaussian Naive Bayes (GNB), Decision Trees, and Gradient Boosting (GBoost). The main objective is to evaluate and compare the performance of these algorithms using a comprehensive dataset sourced from Kaggle, which includes 25,192 data and 42 features. Using metrics such as accuracy, precision, recall, and F1-score, the results show a complex pattern in the strengths and weaknesses of each. Surprisingly, CNN achieved exceptional accuracy, raising questions that require further investigation.