Enhanced Intrusion Detection in Cloud Security by Optimizing Classification Algorithms

The primary threat to network security is detecting intrusive activities over the network. Because network models and bandwidths are rapidly evolving, the effectiveness of Intrusion Detection Systems (IDSs) requires intuitive Artificial Intelligence (AI). Furthermore, studies in recent years have shown that parsing network traffic data into an IDS is a productive method for detecting security breaches. Detecting malicious assaults on any system is the first step in securing it. To detect whether a network traffic is intrusive or normal, classification algorithms support the IDSs in classifying network traffic types. In this paper, we display the comparative results obtained from conducting the CSE-CICIDS2018 dataset base, which represents in information security research the most reliable and up-to-date dataset. Applying the new AI trends to evaluate different feature selections and machine learning techniques. The classification algorithms such as Multilayer Perceptron (MLP), Naïve Bayes, Radial Basis Neural Network (RBF), Support Vector Machine (SVM), k-Nearest Neighbours (KNN), and Decision Tree (J48) algorithms come up to show the unwanted features to be removed from nonbehavioral intrusions and identify the most beneficial combination of qualities for a specific statistical report depending on the accuracy, precision, recall, and ROC curve, leading to the best model used with clarification. Results figured that the best classification algorithm is KNN with an accuracy of 100%.