Search

Mobile traffic classification/ clustering is an important foundation for mobile network traffic management. However, the mobile network traffic data used by different papers were collected from different network environment. In addition, the labels and the flow statistical features of mobile traffic were different from papers. These experimental results couldn't be directly compared. This paper collected the traffic data generated by App based on MobileGT system. The two kinds of labels were built on these data (App level and function level),and two kinds of flow statistical features were independently extracted on these traffic data. This paper comprehensively researched the machine learning techniques on the traffic data with different labels and different flow statistical features. The experimental results show that the uni-direction flow based features are better than bidirection flow based features, random forest and AdaBoost are better on classifying mobile traffic data, and K-means is better on clustering mobile traffic data. [ABSTRACT FROM AUTHOR]

To handle the mobile traffic identification problem, based on multiple performance evaluation metrics, this paper analyzed the performance of K-means and spectral clustering algorithms on the data sets characterized by different feature sets or labeled with different class set, and proposed an ensemble clustering method from the aspects of combining the clustering results on the data sets with different feature sets. Experimental results show that the performance of the same clustering algorithm is different on the data sets with different feature sets or traffic classes, and the ensemble clustering method is able to improve the overall clustering performance. Further, this paper applied the ensemble clustering method on the correlation analysis of mobile Apps, and the results can support the decision on grouping Apps and analyzing user behaviors. [ABSTRACT FROM AUTHOR]

In order to systematically analyze Internet traffic features, this paper researched the taxonomy of traffic features based on the objects used for building traffic features,and overviewed the related traffic features in each category. It further analyzed the stability of traff ic features, the impact of packet sampling, network environment and obfuscating on traffic features. And then it compared the different kinds of traffic features from classification accuracy .stability, timeliness and granularity. Finally, this paper concluded future works about traffic features. [ABSTRACT FROM AUTHOR]