Quantum Data Traffic Analysis for Intrusion Detection System
Anshul Harish Khatri, Vaibhav Gadag, Simrat Singh, Sandeep Kumar Satapathy, Debahuti Mishra · 2023
An IDS (Intrusion Detection System) is a device or software that inspects all network traffic and notifies the user or administrator if unauthorized attempts or access have occurred. Many assaults, on the other hand, make it censorious for security of networks since they permit you to observe and act against malicious traffic. These attacks primarily are dependent on various techniques pertaining to social engineering which the attackers can exploit to get user's data credentials and get access to target's network and active assets. Despite its potential and the fact that it is the focus of many academics, its applicability to real-world applications has been impeded by system complexity, since these systems require extensive testing, assessment, and tuning before deployment. The most idealistic way for testing and assessment is to run these systems on actually labeled network traces with a full and large set of intrusions and anomalous behaviors. The proposed approach detects such attacks by analyzing the transmitted packet properties, evaluating the models, and visualizing the output. A dataset of 1048575 rows and over 80 columns is used in this study. Several machine learning algorithms such as Random Forest (RF), Decision Tree (DT), Ada Boost Classifier (ADA), Ridge Classifier (Ridge), Logistic Regression (LR), SVM-Linear Kernel (SVM), Naive Bayes (NB), and Quadratic Discriminant Analysis (QDA), with the help of Auto ML were used in the model creation and testing. The results obtained indicate a reasonable accuracy rate. For each data file, when applied autoML generated the best-fit algorithm with unique metrics with accuracies as high as 99%. These saves were then used for the prediction and generation of an output CSV which was mapped onto various plots using the AutoViz library.