Intrusion Detection and Attack Classification using an Ensemble Approach

Mantek Singh, Ridam Arora

Abstract:

he challenges to ensure safe and trusted
communication of information between various
organizations have increased multifold in recent past.
Intrusion Detection Systems such as firewall, message
encryption and other approaches are being employed with
partial success, however the risks and chances of malicious
intrusions are still posing a threat. We are proposing to
make use of recent advancements in the field of machine
learning to develop an intrusion detection system. In our
work, the machine learning classifiers namely, random
forest, decision table, multi-layer perceptron and naive
bayes were used in an ensemble model showing a significant
improvement in the overall accuracy. The proposed
approach was implemented using a bench-marking dataset
from KDDCup.
Key Words: Ensemble learning, machine learning,
intrusion detection