Predicting Distributed Denial Of Service Attack With Mining Based Approach

Abstract

Wireless sensor network provides resources as per requirement of the user. WSN consists of sensors arranged in sequence for sending and receiving signals. WSN is hampered with the attacks such as distributed denial of service, WORM hole attack etc. This work present ensemble-based approach for detecting DDOS attack caused by malicious users. The impact of DDOS attack on the WSN along with need to tackle DDOS attack is discussed. For accomplishing the detection process, ensembles based voting classifiers is designed. Ensemble based algorithms used for demonstrating the DDOS attack includes Regression mechanism that could be linear or nonlinear in nature, Hyperplane dependent SVM that is also termed as multi support vector machine, random forest that is basically used to select branched that have optimal probability of being attackers., naïve bayes and K-nearest neighbour approach for forming cluster of
nearest attribute nodes. The classification accuracy of combined classifier is better as demonstrated within the result section.