Effective and efficient approach in IoT Botnet detection

Abstract

Internet of Things (IoT) technology presents an advantage to daily life, but this advantage is not a guarantee of security. This is because cyber-attacks, such as botnets, remain a threat to the user. Detection systems are one of the alternatives to maintain the security of IoT network. A reliable detection system should effectively detect botnets with high accuracy levels and low positive rate. It should be efficient to perform detection quickly. However, data generated by IoT networks have high dimensions and high scalability, so they need to be minimized. In network security analysis process, high-dimension data pose challenges, such as the dimension curse correlation between different dimensions, which causes features that are hard to define, datasets that are mostly unordered, cluster combination, and exponential growth. In this study, we applied feature reduction using the Linear Discriminant Analysis (LDA) method to minimize features on the IoT network to detect botnet. The reduction process is carried out on the N-BaIoT dataset which has 115 features reduced to 2 features. Performing feature reduction with detection systems has become more effective and efficient. Experimental result showed that the application of LDA combined with machine learning on the classification Decision Tree method was able to detect with accuracy that reached 100% in 98.58s with only two features.