Detecting network traffic irregularity with combination of k-means and generalized likelihood ratio
| Autor: | Emmanuel Boachie, Okyere Stephen, Lichunlin Li |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Matching (graph theory)
Computer science Network packet Frame (networking) k-means clustering Centroid 020206 networking & telecommunications 02 engineering and technology Residual 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Cluster analysis Host (network) Algorithm |
| Zdroj: | ICTCE |
| DOI: | 10.1145/3145777.3145796 |
| Popis: | Data extraction techniques have made it possible to analyze big set applying rules and patterns. In area of networking, data extraction techniques can be possibly applied to check invasion, attacks or anomalies so far as recording network checking data on a host is concerned. Furthermore, there are drawbacks when it comes to implementation of algorithm of extracting techniques to check anomalies. Most network checking systems requires application of K-means algorithm for modeling due to its efficient and effective centroid cluster. Many researchers have identified the problem and suggest method to solve without using K-means algorithm for modeling. This paper crosschecks the Network Traffic irregularity Detections and recommends different algorithm to check traffic failure of frame/packet and flow data captured in a network. The researcher presents Clustering, K-means Generalized Likelihood Ratio (KmGLR) algorithm which is an effective and efficient to check breakdown end. KmGLR algorithm combines K-means clustering with Generalized Likelihood Ratio algorithm. Training data containing unlabeled stream records are separated in clusters after that, KmGLR algorithm is deployed to calculate the residual error and check network irregularity traffic. The matching cluster centroids are used as patterns to compute efficient deployment of KmGLR algorithm for detection of anomalies. Simulation test has affirmed that the KmGLR algorithm performs proper function in error and irregularity checking as compared to others. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|