Efficient Anomaly Intrusion Detection Using Hybrid Probabilistic Techniques in Wireless Ad Hoc Network

A wireless ad-hoc network includes huge number of mobile nodes that structure temporary networks. Due to the dynamic nature of wireless ad-hoc network, security and efficient intrusion detection system (IDS) is a challenging task to detect the intruder nodes. The classification algorithm is used to detect the intrusions in an efficient manner. However, the network is characterized by high mobility they also introduce many vulnerabilities that increase their accurate detection risks. The optimization technique is performed to attain effective model for intrusion detection. But, the IDS continuously use additional resources to monitoring intruder activity in the network. In order to overcome the above issues in wireless ad-hoc network, Simulated Annealing based Naive Bayes Classifier (SA-NBC) technique is proposed for Anomaly Intrusion Detection. An anomaly-based intrusion detection system is used to detect the network intrusions and monitoring network activities in an exact manner. At first, the optimal features are chosen for classifying and detecting the intrusion by means of Simulated Annealing (SA) method when performing packet transmission. Based on these selected features, the accuracy and efficiency of traffic pattern analysis is improved using intrusion detection. Next, the Naive Bayes classifier is employed to classify the attack depends on features to identify the malicious behavior accurately from normal node in a testing environment by using the Bayes theorem. This in turns, the network traffic is minimized and increases the accuracy of anomaly intrusion detection. The SA-NBC technique conducts the simulations work on parameters such as anomaly intrusion detection accuracy, execution time, and throughput. The simulation results demonstrate that the SA-NBC technique is able to improve the accuracy of intrusion detection and also improves the throughput when compared to state-of-the-art works.