Machine learning based internet traffic recognition with statistical approach

Jaiswal, Rupesh; Lokhande, Shashikant

doi:10.1109/indcon.2013.6726074

Cited by 22 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And due to the discriminative ability of the supervised learning, we would select common and comprehensive features as input to provide sufficient statistics for different tasks of classifications. Thus, we concluded from the previous studies [5]- [7] and decided to implement totally 16 features promising the unbias from feature selection (which has been evaluated by matrix factorization methods, such as variational bayesian matrix factorization [34]): transport layer source port, destination port, number of packets with PUSH flag, ratio of upload and download, the first quartile of inter-arrival time and packet size, the statistical properties of the inter-arrival times and packet sizes (the minimum, the maximum, the mean, the variance and the rooted mean square).…”

Section: Datasets and Feature Selectionsupporting

confidence: 51%

“…Our work was motivated by the fact that the demands will always be changing and that most of the previous works have shown that some features can be used to discriminate different tasks of flows. For instance, the features for classifying either elephant or mice flows and that for identifying the applications of flow, are surprisingly similar [5]- [7]. This alludes to that there might be some common knowledge for different traffic classification tasks and the mapping functions from the selected features to the different flow properties that we need to predict are of close relations.…”

mentioning

confidence: 96%

See 1 more Smart Citation

Common Knowledge Based and One-Shot Learning Enabled Multi-Task Traffic Classification

et al. 2019

View full text Add to dashboard Cite

Deep neural networks have been used for traffic classifications and promising results have been obtained. However, most of the previous work confined to one specific task of the classification, where restricts the classifier potential performance and application areas. The traffic flow can be labeled from a different perspective which might help to improve the accuracy of classifier by exploring more meaningful latent features. In addition, deep neural network (DNN)-based model is hard to adapt the changes in new classification demand, because of training such a new model costing not only many computing resources but also lots of labeled data. For this purpose, we proposed a multi-output DNN model simultaneously learning multi-task traffic classifications. In this model, the common knowledge of traffic is exploited by the synergy among the tasks and improves the performance of each task separately. Also, it is showed that this structure shares the potential of meeting new demands in the future and meanwhile being able to achieve the classification with advanced speed and fair accuracy. One-shot learning, which refers to the learning process with scarce data, is also explored and our approach shows notable performance. INDEX TERMS Traffic classification, multi-task model, transfer learning, one-shot learning. I. INTRODUCTION

show abstract

Section: Datasets and Feature Selectionsupporting

confidence: 51%

mentioning

confidence: 96%

Common Knowledge Based and One-Shot Learning Enabled Multi-Task Traffic Classification

et al. 2019

View full text Add to dashboard Cite

show abstract

“…It turns out that revealing the basic patterns of the traffic data at the base station level, can represent an important solution for those providers that face network congestion problems or over-provisioning of the link capacity. Then, we investigated previous work on traffic data classification and identification in [2], [3] and [4]. They deal with different methods, from statistical to principal component analysis approaches, obtaining results with high accuracy.…”

Section: Related Workmentioning

confidence: 99%

“…Unfortunately, the effectiveness of such packet inspection techniques is diminishing because they rely on two related assumptions: 1) third parties unaffiliated are able to inspect each IP packet's payload; 2) the classifier knows the syntax of each application's packet payloads. Newer approaches classify internet traffic by recognising statistical patterns in externally observable attributes [3] (such as typical packet lengths and inter-packet arrival times). They aim to cluster IP traffic flows into groups that have similar traffic patterns, or classify one or more applications.…”

Section: Introductionmentioning

confidence: 99%

Identification of tidal-traffic patterns in metro-area mobile networks via Matrix Factorization based model

Troia

Gao

Alvizu

et al. 2017

2017 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops)

View full text Add to dashboard Cite

“…The evaluation results identified the effectiveness of such structures. After introducing four types of machine learning techniques, Jaiswal and Lokhande [24] demonstrated a classification scheme for network traffic. Both the new captured dataset and standard dataset are adopted for investigation purpose.…”

Section: Related Workmentioning

confidence: 99%

Designing Efficient Mobile Gateway for Impoverished Regions Based on Self-Learning Strategy

Zhou

Shi

Song

et al. 2016

Mobile Information Systems

View full text Add to dashboard Cite

Multiple wireless access technology had tremendously extended the coverage of Internet from developed cities to impoverished regions. As the significant network equipment, mobile gateway had been assigned more and more tasks. However, the efficiency issues are still challenging. Motivated by the latest progress in self-learning, we proposed a new scheme for mobile gateway design. Firstly, the network architecture and supported new features of mobile gateway are discussed to provide a macroscopic impression. Secondly, the application scenarios and functional modules are explored from network and user oriented plane’s perspectives. The implementation of intraplane and interplane interactions is carefully illustrated. Thirdly, the transitions of multiple states are analyzed based on possible network context. Fourthly, three typical self-learning schemes are briefly examined and two representative algorithms are adopted. Necessary procedures for deployment are also demonstrated via pseudocodes. Then the interactive process in both network and user oriented cases is investigated. In order to validate the performance, a comprehensive topology with background traffic is established. Simulation results reveal that our hybrid scheme could overcome single parameter based scheme in most of the scenarios.

show abstract

Machine learning based internet traffic recognition with statistical approach

Cited by 22 publications

References 18 publications

Common Knowledge Based and One-Shot Learning Enabled Multi-Task Traffic Classification

Common Knowledge Based and One-Shot Learning Enabled Multi-Task Traffic Classification

Identification of tidal-traffic patterns in metro-area mobile networks via Matrix Factorization based model

Designing Efficient Mobile Gateway for Impoverished Regions Based on Self-Learning Strategy

Contact Info

Product

Resources

About