Upgrade Better Accuracy by Analyzing the Behavior of Classifying Algorithm to Detect Intrusion in Network Traffic

One of the biggest obstacles of improving network security is accurately identifying and stopping intrusions. Using the Friday-WorkingHours-Afternoon-DDos.pcap_ISCX11 dataset, this project aims to determine which algorithm has the best accuracy in detecting intrusions, specifically Distributed Denial of Service (DDoS) attacks. The dataset was selected because it is applicable to real-world situations. A comparative analysis is carried out to assess how different classifying algorithms perform in relation to intrusion detection. The project entails a great deal of algorithm experimentation, parameter tuning, and painstaking performance assessment. Finding the algorithm that performs best at handling the unique patterns in the dataset is the goal, and it will eventually help to improve intrusion detection systems. The purpose of this research's finding is to offer insightful information about which intrusion detection algorithm works best for the given dataset. Determining the algorithm with the best accuracy becomes critical to improving network security as cyber threats keep evolving. The results have applications for cybersecurity experts looking to implement strong intrusion detection systems, and they can guide the creation of new algorithmic techniques to counter new threats in the future. The most performing algorithm is CatBoost Classifier and XGBoost Classifier with 99.99% accuracy.