DETECTION OF IOT-BASED CYBER ATTACKS USING MACHINE LEARNING TECHNIQUES

Abstract

The Internet brings Things together a huge number of gadgets that can communicate with one another with minimal user intervention. IoT devices are progressively deployed day-by-day life. Many of these devices are vulnerable due to insecurity, execution, and design. Cyber-attacks are currently posing a significant threat to IoT security. These risks have the potential to steal sensitive information, resulting in financial and reputational harm. IoT devices can now stay connected for longer periods without the need for human interaction. This necessitates the development of intelligent network-based security solutions, such as machine learning systems. Although several studies have examined the utilization of Machine Learning arrangements in attack discovery issues in recent years. The desire to add the field by assessing various ML methods that utilized to distinguish IoT network attacks fast and effectively in this study. A unique dataset, KDD99-IoT, is utilized to assess different discovery calculations. This dataset consists of four types of attacks DoS, Probe, U2L and U2R. In the execution stage, diverse AI calculations are used and the greater part of them accomplished superior. New provisions were removed from the KDD-IoT dataset throughout the execution and contrasted and concentrate the writing, and the new elements provided improved outcomes. In this research the three algorithms were used to detect Cyber-attacks. KNN, Logistic Regression and K-Means Clustering were used, all of these algorithms the Logistic Regression Provide best result 98% accuracy and KNN 97% accuracy. K-mean algorithm are used to make Clusters of the Attacks.