|
Title:
|
AN INTELLIGENT HYBRID DECISION APPROACH WITH FEATURE SELECTION FOR ANOMALY NETWORK INTRUSION DETECTION SYSTEM |
|
Author(s):
|
Jamal Hussain, Samuel Lalmuanawma |
|
ISBN:
|
978-989-8533-29-6 |
|
Editors:
|
Piet Kommers, Tomayess Issa, Dian-Fu Chang and Pedro Isaías |
|
Year:
|
2014 |
|
Edition:
|
Single |
|
Keywords:
|
Network Intrusion Detection, Feature selection, Machine learning, Decision tree, Adaboost, Classification. |
|
Type:
|
Full Paper |
|
First Page:
|
3 |
|
Last Page:
|
10 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The network intrusion detection system (NIDS) assists to prevent our network systems against sophisticated attacks and malwares. However, traditional technique for network prevention system using individual classifier fails to completely protect network systems due to rapid evolution of different new pattern of attack. This paper presents a new hybrid approach for NIDS using ensemble method, combining Adaboost (AB) algorithm with C4.5 decision tree (DT) classification, DT can find the optimal features selection to ameliorate the accuracy of anomaly intrusion detection, by automatically adjusting the optimal parameter settings for the proposed model. After applying features selection classifier using unsupervised wrapper method over the whole dataset, the output data is again applied to a classifier for further classification. To carry out our experiment we used k-fold cross validation over two-class (i.e., normal and anomaly) classification strategy where k is set to 10. Our proposed new hybrid NIDS technique was evaluated with the NSL-KDD datasets, which is a customized and enhanced edition of KDD99 data sets developed by DARPA. The proposed new model outperforms other existing conventional approaches for both individual and hybrid classification, simulation results describes that the proposed new model is reliable in detecting anomaly intrusion detection system. |
|
|
|
|
|
|
