Identify Promising Classifiers for Each Type of Attack Class

Conference: Fourth International Conference on Advances in Computer Science and Application
Author(s): Yogesh Kumar, Indu Bala Year: 2015
Grenze ID: 02.CSA.2015.4.8 Page: 8-17

Abstract

Most current offline intrusion detection systems are focused on unsupervised and supervised machine learning approaches. Due to the variety of network behaviors and the rapid development of attack fashions, it is necessary to find the best machine-learning-based intrusion detection algorithms with high attack rates. There are many Classification mechanisms used in data mining that would help with intrusion detection system such as Naive Bayes Algorithm, IBK, J48, Random forest, AttributeSelectedClassifier(ASC), ClassificationviaRegres-sion(CVR), Decision stump, REPTree, Random Tree, Filtered classifier, RandomCommitee, JRip, HoeffdingTree. This paper presents a comparison of 14 classification techniques based on the performance measures TP rate,FP rate,Precision, Fmeasure, ROC area. The goal of this research is to enumerate the high attack rates techniques from above fourteen analyzed algorithms under a given data set and provide a fruitful comparison result.

<< BACK

CSA - 2015