<![CDATA[IoT devices have played an important role in driving DDoS attacks, and are a threat to IoT networks. One of them is the ICMP Flood attack. To overcome attacks on IoT, one of them uses an Intrusion Detection System (IDS). However, on the other hand, IDS has challenges in handling the complexity of high-dimensional data. One of the suggested solutions to overcome the problem of data dimensions is the use of feature selection techniques. The Forward Selection feature selection technique is used to eliminate irrelevant features. This study compares the performance of the Random Forest and SVM algorithms. For experimental purposes, this study used the CICIoT2023 dataset, which represents IoT traffic. The use of Forward Selection obtained 11 selected features that will be used in the machine learning process using the Random Forest and SVM methods. Feature selection affects the computation time or processing time, because the fewer features used, the more the system's workload in carrying out the classification process. The test results show that the use of feature selection improves the performance of random forest with an accuracy of 100%. Meanwhile, the SVM model gets better accuracy by using feature selection with the highest accuracy of 99.4508% in the supplied test set test.]]>
Copyrights © 2025
