Feature selection for intrusion detection: an evolutionary wrapper approach

Hofmann, Alexander; Horeis, T.; Sick, Bernhard

doi:10.1109/ijcnn.2004.1380189

Cited by 32 publications

(18 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This data was originally discussed in ( [4,5,13]) and the actual data used in this paper was also discussed in ( [6][7][8]). The Sick data was originally examined in [11]. The parameter ν is set to .5 for every experiment.…”

Section: Resultsmentioning

confidence: 99%

Some Properties of the Gaussian Kernel for One Class Learning

Evangelista

Embrechts

Szymański

2007

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper proposes a novel approach for directly tuning the gaussian kernel matrix for one class learning. The popular gaussian kernel includes a free parameter, σ, that requires tuning typically performed through validation. The value of this parameter impacts model performance significantly. This paper explores an automated method for tuning this kernel based upon a hill climbing optimization of statistics obtained from the kernel matrix.

show abstract

Section: Resultsmentioning

confidence: 99%

Some Properties of the Gaussian Kernel for One Class Learning

Evangelista

Embrechts

Szymański

2007

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Feature selection based on rough set was introduced by [12]. (Stein et al, 2005) [13] utilized decision tree classifiers with genetic algorithm in intrusion detection approach while (Hofman et al,2004) [16] gave evolutionary wrapper approach (combining radial basis function with genetic algorithm). Genetic algorithm along with naïve bayes classification was used by (Lee et al, 2006) [11] to generate optimized results for classification on KDD99 dataset.…”

Section: Related Workmentioning

confidence: 99%

“…To ensure that for each rule, the corresponding class selected is correct, following formula is applied- (16) Hence the main objective is to maximize T and thus cover approximately all the samples of training dataset. Therefore the feasibility of obtaining the best rule set will be maximized which will improve the classification of attacks in testing dataset as more features will be available for classification.…”

Section: Fuzzy Rule Based Systemmentioning

confidence: 99%

Hybrid Genetic Fuzzy Rule Based Inference Engine to Detect Intrusion in Networks

Chadha

Jain

2015

Intelligent Distributed Computing

View full text Add to dashboard Cite

Abstract.With the drastic increase in internet usage, various categories of attacks have also evolved. Conventional intrusion detection techniques to counter these attacks have failed and thus substantial systems are needed to eliminate these attacks before they inflict huge damage. With the ability of computational intelligence systems to adapt, exhibit fault tolerance, high computational speed and error resilience against noisy information, a hybrid genetic fuzzy rule based inference engine has been designed in this paper. The fuzzy logic constructs precise and flexible patterns while the genetic algorithm based on evolutionary computation helps in attaining an optimal solution, thus their collaboration will increase the robustness of intrusion detection system. The proposed network intrusion detection system will be able to classify normal behavior as well as anomalies in the network. Detailed analysis has been done on DARPA-KDD99 dataset to specify the behavior of each connection.

show abstract

“…For an analysis of the selected features for the present data set, in particular for Guest, see [75]. A lot of possible extensions or improvements of the EA are discussed in [70].…”

Section: Automated Feature Selection and Structure Optimizationmentioning

confidence: 99%