Dynamic topology and limited resources are major limitations that make intrusion detection in mobile adhoc network (MANET) a difficult task. In recent years, several anomaly detection techniques were proposed to detect malicious nodes using static and dynamic baseline profiles, which depict normal MANET behaviors.
In this research, we investigated different baseline profile methods and conducted a set of experiments to evaluate their effectiveness and efficiency for anomaly detection in MANETs using C-means clustering technique. The results indicated that a static baseline profile delivers similar results to other baseline profile methods. However, it requires the least resource usage while a dynamic baseline profile method requires the most resource usage of all the baseline models.