Hybrid data mining approach for intrusion detection using modified AODV algorithm

Ad-hoc networks are collection of self-organized, autonomous nodes (routers) and capable to communicate directly with neighbors' node through broadcast packet in within transmission range. Ad hoc network has a primary concern to provide protected communication between mobile nodes. In transmission channel, nodes can be a malicious node and compromise or breaches service, which makes security extremely challenging. Our major focus is to discuss the feasibility of monitoring the nodes of different networks, and analyze it for providing better security. Data mining techniques used to classify for large aggregate data according classification rules and patterns, to detect or identify malicious node. In this paper Idea is based on k-Mediods clustering algorithm to form cluster with high detection rate based on intrusion behavior or normal behavior.