Skip to content

The Classification with Machine Learning of User Application Activity on Network Traffic

Notifications You must be signed in to change notification settings

elifyener/network-traffic-classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

The Classification with Machine Learning of User ApplicationActivity on Network Traffic

Collaborators: Elif Yener | İrem Aysel İpek | Muhammet Ali Karagüzel

In this study, machine learning techniques have been applied to classify network traffic using network flow features and application tags. The classification techniques are a data mining technique for using an estimate that the data belong to which category. In this study, the assignment of packets in the network flow traffic of a tagged data set to predetermined application groups is provided by classification algorithms.

The supervised classification algorithms to used in this study are k-Nearest Neighbor, Decision Tree, Random Forest, Gaussian Naive Bayes, Support Vector Machine, and Logistic Regression. The strengthening algorithm we use is XGBoost.

Weka software, which is a data mining application, was used to balance the flow sample inequality between the results obtained from the unbalanced data and the applications. Applications with a high ratio have been reduced. The lower ones have synthetically enhanced thanks to the Synthetic Minority Over-sampling(SMOTE) filter.

To have a high accuracy value used classification algorithms, selecting proper hyperparameter is important. This process was done with the Random Search or Grid Search algorithm, because countless trial and error methods are necessary, to find the best selection of parameters. The K-Fold Cross Validation algorithm was used to understand whether the high performance on the model provided by the hyperparameters used in the algorithms is random or not.

About

The Classification with Machine Learning of User Application Activity on Network Traffic

Topics

Resources

Stars

Watchers

Forks

Languages