Comparative Analysis of Random Forest and Decision Tree for Gray Hole Attack Detection in Wireless Ad Hoc Networks
T. J. Nagalakshmi, R. Samar Kumar · 2024
Developing two intrusion detection systems (IDS) to identify grey hole attacks in wireless ad hoc networks is the goal of this project. For this, the Random Forest (RF) and Decision Tree algorithms were used. The NS2 tool was used to create a dataset with 90 nodes and 12 network layer characteristics. Using the SPSS programme, the IDS models were constructed and assessed, using 19 samples in each group. The findings show good performance (accuracy: 0.003; detection rate: 0.041) at a significance level of p<0.05. Whereas the IDS using Decision Tree obtained 99.49% accuracy and 99.57% detection rate, the IDS using RF reached 99.89% accuracy and 97.94% detection rate. This indicates that when it comes to identifying grey hole attacks, the RF-based IDS performs noticeably better than the Decision Tree-based IDS.