HYBRID INTRUSION DETECTION METHOD BASED ON IMPROVED ADABOOST AND ENHANCED SVM FOR ANOMALY DETECTION IN WIRELESS SENSOR NETWORKS

The utilisation of Wireless Sensor Networks is quickly rising due to the fast progress of wireless sensor technologies. Due to limited resources, infrastructureless nature, and other factors, it faces major security difficulties. This study describes a hybrid IDS based on an improved AdaBoost and Enhanced SVM strategy for detecting network intrusions and monitoring node activity while classifying it as normal or abnormal. AdaBoost is used in combination with an SVM classifier to identify and classify intrusions. The suggested IDS considerably enhanced the network performance by recognising and eliminating malicious nodes from the network and avoiding DoS and sinkhole attacks. Results oproved that it performes better than other state of art methods in terms of transmission delay, detection rate, energy consumption, packet delivery rate. It also has the advantages of a simple structure and quick computation times.