PDF Publication Title:
Text from PDF Page: 031
3.2.2 Traffic prediction Traffic prediction is a classic problem in networking which has been usually tackle with time- series prediction models [8][84] such as: Hidden Markov Model, Exponential Smoothing, AutoRegressive Integrated Moving-Average (ARIMA), ARIMA with eXogenous covariates (ARIMAX), etc... [85][86] The problem of traffic prediction consists in forecasting the traffic volume or connectivity status of a communication flow in the short, medium- or long-term future. Besides the above-mentioned time-series models the usual ML models applied to this problem have been: Neural Networks, SVM and Recurrent Neural Networks-LSTM (more recently) [8][86] All the experiments in this area are performed with proprietary data as it is the case for the research performed in this thesis. The prediction of network traffic using deep learning does not have much literature. In the field of the prediction of network activity/traffic, the use of other machine learning methods is more frequent, mainly time-series methods [87][88][89]. Considering some of the few examples of application of deep learning to traffic forecasting: in [90] a deep learning network (CNN, RNN) is used for mobile traffic forecasting using spatiotemporal features; and for a similar problem, in [91] a combination of LSTM and Stacked Autoencoder networks is applied. The following table presents a summary of the main works related to the research carried out for this thesis. It provides a reference to the document, the data set used and the scope of the work. Objective/Area Ref. Dataset Scope Traffic prediction - Supervised [43] Simulated network data reliability for wireless sensor networks, using two parameters to make the prediction number of sensors nodes and transmission range. It is a binary prediction but of a very different nature to this thesis works. - Prediction using logistic regression of network [44] Real WLAN traffic traces - Using SVM to forecast traffic in WLANs. They study the issues of one-step-ahead prediction and multi-step-ahead prediction - Prediction of continuous values (with MSE and NMSE) which makes them impossible to compare with the results of this thesis work (discrete prediction). - SVM presents better results than MLP, ARIMA and Fractional ARIMA. They provide the Mean Square Error for different time-steps predictions. Doctoral Thesis: Novel applications of Machine Learning to NTAP - 29PDF 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 (Standard Web Page)