Paper Title
Using Dimensionality Reduction Technique for Valuable IoT Botnet Data Extraction

As the attention and demand around IoT systems is getting higher, it became a target to cyber attackers. A very famous way to attack IoT systems is through spreading and gaining control over IoT botnets. IoT botnet can be a computer or any IoT device such as sensors, monitors, cameras...etc. so they can be remotely accessed and used for malicious actions. To address this issue, many researchers have developed a machine learning solution to detect and prevent such attacks. However, all these solutions start with selecting good dataset to extract relevant botnets features for classification purposes. The complexity of such classifier can be greatly reduced if the numbers of attributes in a data set are reduced. The objective of this paper is to incorporate dimensionality reduction technique to in order to produce a better data extraction with an optimal number of features. To prove the effectiveness of the proposed technique, a comparative study will be carried out. The study will start by developing a simple machine learning classifier to detect IoT botnets and the resulted system will be tested for performance on both original dataset and the optimized version with reduced number of features using dimensionality reduction techniques. Keywords - Internet of Things (IOT), Cybersecurity, Dimensionality Reduction, Machine Learning, Feature Extraction