In machine learning, balanced datasets play a crucial role in the bias observed towards classification and prediction. The CSE-CIC IDS datasets published in 2017 and 2018 have both attracted considerable scholarly attention towards research in intrusion detection systems. Recent work published using this dataset indicates little attention paid to the imbalance of the dataset. The study presented in this paper sets out to explore the degree to which imbalance has been treated and provide a taxonomy of the machine learning approaches developed using these datasets. A survey of published works related to these datasets was done to deliver a combined qualitative and quantitative methodological approach for our analysis towards deriving a taxonomy.
The research presented here confirms that the impact of bias due to the imbalance datasets is rarely addressed. This data supports further research and development of supervised machine learning techniques that reduce bias in classification or prediction due to these imbalance datasets. This study's experiment is to train the model using the train, and test split function from sci-kit learn library on the CSE-CIC-IDS2018. The system needs to be trained by a learning algorithm to accomplish this. There are many machine learning algorithms available and presented by the literature. Among which there are three types of classification based Supervised ML techniques which are used in our study: 1) KNN, 2) Random Forest (RF) and 3) Logistic Regression (LR). This experiment also determines how each of the dataset's 67 preprocessed features affects the ML model's performance. Feature drop selection is performed in two ways, independent and group drop. Experimental results generate the threshold values for each classifier and performance metric values such as accuracy, precision, recall, and F1-score. Also, results are generated from the comparison of manual feature drop methods. A good amount of drop is noticed in the group for most of the classifiers.
Library of Congress Subject Headings
Machine learning; Data sets--Quality control; Computer security--Data processing
Networking and System Administration (MS)
Ravikumar, Dharshini, "Towards Enhancement of Machine Learning Techniques Using CSE-CIC-IDS2018 Cybersecurity Dataset" (2021). Thesis. Rochester Institute of Technology. Accessed from