PDF Publication Title:
Text from PDF Page: 126
Fig. 11. Classification performance metrics (aggregated) vs. time-series length for architecture CNN+RNN-2a Therefore, it seems clear that although a minimum number of packets is important, a large number of packets (that is architecture-dependent) is not necessary. In general, between 5 and 15 packets are enough to achieve excellent detection results. V . CONCLUSION This work is a contribution to improve the available alternatives and capabilities of NTC in current network monitoring systems; being specially targeted to IoT networks, where traffic classification is highly required [1, 2]. As far as we know, there is no previous application of the RNN and CNN deep learning models to an NTC problem. Therefore, the work presented in this paper is original in essence. This work provides a thorough analysis of the possibilities provided by deep learning models to NTC. It shows the performance of RNN and CNN models and a combination of them. It demonstrates that a CNN can be successfully applied to NTC classification, giving an easy way to extend the image-processing paradigm of CNN to a vector time-series data (in a similar way to previous extensions to text and audio processing [34, 35]). A model based on a particular combination of CNN plus RNN gives the best detection results, being these results better than other published works with alternative techniques. The impact of selected features is demonstrated, and also that it is not necessary to process a large number of packets per flow to have excellent results: any number of packets higher than 5-15 (a number which is architecture-dependent) gives similar results. The proposed method is robust and gives excellent F1 detection scores under a highly unbalanced dataset with over 100 different classification labels. It works with a very small number of features and does not require feature engineering. To train the models we have made use of high-level header-based data extracted from the packets. It is not required to rely on IP addresses or payload data, which are probably confidential or encrypted. A simple RNN model provides already very good results, but it is interesting to appreciate that these results improve when the RNN model is combined with a previous CNN model. Being it possible to improve results with the inclusion of a CNN shows how the initial intuition that allowed us to assimilate the vector time-series extracted from network packets’ Doctoral Thesis: Novel applications of Machine Learning to NTAP - 124PDF Image | Novel applications of Machine Learning to Network Traffic Analysis
PDF Search Title:
Novel applications of Machine Learning to Network Traffic AnalysisOriginal File Name Searched:
456453_1175348.pdfDIY PDF Search: Google It | Yahoo | Bing
Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info
Cruising Review Topics and Articles More Info
Software based on Filemaker for the travel industry More Info
The Burgenstock Resort: Reviews on CruisingReview website... More Info
Resort Reviews: World Class resorts... More Info
The Riffelalp Resort: Reviews on CruisingReview website... More Info
| CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com | RSS | AMP |