Abstract-Network security is one of the major concerns of the modern era. With the rapid development and massive usage of internet over the past decade, the vulnerabilities of network security have become an important issue. Intrusion detection system is used to identify unauthorized access and unusual attacks over the secured networks. Over the past years, many studies have been conducted on the intrusion detection system. However, in order to understand the current status of implementation of machine learning techniques for solving the intrusion detection problems this survey paper enlisted the 49 related studies in the time frame between 2009 and 2014 focusing on the architecture of the single, hybrid and ensemble classifier design. This survey paper also includes a statistical comparison of classifier algorithms, datasets being used and some other experimental setups as well as consideration of feature selection step.

No abstract

Increasing internet usage and connectivity demands a network intrusion detection system combating cynical network attacks. Data mining therefore is a popular technique used by intrusion detection system to prevent the network attacks and classify the network events as either normal or attack. Our research study presents a wrapper approach for intrusion detection. In this framework Feature selection technique eliminate the irrelevant features to reduce the time complexity and build a better model to predict the result with a greater accuracy and Bayesian network works as a base classifier to predict the types of attack. Our experiment shows that the proposed framework exhibits a superior overall performance in terms of accuracy which is 98.2653 , error rate of 1.73 and keeps the false positive rate at a lower rate of 0.007. Our model performed better than other leading state-of-the-arts models such as KNN, Boosted DT, Hidden NB and Markov chain. The NSL-KDD is used as benchmark data set with Weka library functions in the experimental setup.

No abstract

The feature selection approach provides improved prediction and minimizes the computation time. Due to the higher numbers of features the understanding of the data in pattern recognition becomes difficult sometimes. That's why researchers have used different feature selection techniques with the single classifiers in their intrusion detection system to build up a model which gives a better accuracy and prediction performance. In this paper, we provide a comparative analysis with the feature selection approach in WEKA machine learning tool using the J48 classifier. The research work show the comparison of the performance of single J48 classifier with filter methods. The prediction performance may differ marginally in some cases but with the removal of irrelevant features time complexity can be easily ignored and a better prediction rate is guaranteed.