Nowadays due to serious threats in wireless sensor networks has led the organizations to take serious measures to protect their data against potential attacks, malware, software or platform susceptibilities, etc. There are different kinds of vulnerable security attacks which are mainly classified into Active and Passive. Active attacks are those which intents to corrupt the data or network and on the other hand, Passive attacks intrudes to monitor the traffic of the data. If proper security plan is not taken into account, the sensor networks might be vulnerable to any kinds of attacks. There are many researches taken place since decades to detect the presence of attacks in a network and take appropriate measures to control it before it hinders the data or network. So here the aim is to develop a Machine Learning Model that detects whether a wireless sensor network is poisoned with an attack or not. If an attack is encountered, this model will have to find the type of the active attack. In order to get the real data, we have created our own dataset by performing those attacks in a confined network and collecting the data, cleaning them and doing appropriate preprocessing. Followed by this we have used those data and developed a machine learning model that predict whether an attack occurred in a network. The accuracy of the model has considerably achieved desired output (97 percent accuracy).
Volume 11 | Issue 12
Pages: 82-89
DOI: 10.5373/JARDCS/V11I12/20193215