The Impact of Feature Selection on Malware Classification Using Chi-Square and Machine Learning
Areeg Fahad Rasheed, M. Zarkoosh, Sana Sabah Al-Azzawi
Abstract
The Internet of Things (IoT) is a network of physical objects, automobiles, household appliances, and other items that are integrated with sensors, software, and connections to gather and share data via the Internet. The rapid proliferation of the Internet of Things devices has ushered in a wave of new security challenges, specifically in the realm of malware detection and these challenges necessitate innovative solutions. Consequently, the primary objective of this study is to develop an advanced malware detection system leveraging machine learning algorithms in tandem with natural language processing techniques like tokenization and vectorization, in addition to a feature selection method named Chi-Square. The proposed method is tested using the IoTPot dataset and compared with recent research in the field, where it outperformed the current work with respect to the accuracy, F1-score, recall, and precision. Furthermore, the proposed method was compared with time-based consulting and demonstrated superior performance with NLP and Chi-square than without, making it more suitable for resources constrained by such IoT systems. We also provide the code for the proposed method to foster transparency. <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> . <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> https://github.com/AREEG94FAHAD/chisqaure