Scribd
Upload
32 views1 page

PHD Proposal Title: Internet Traffic Forecasting Using Data Mining Techniques

This PhD proposal aims to develop advanced data mining techniques like neural networks and support vector machines to forecast internet traffic using limited information. The objectives are to forecast traffic for specific network topologies, aggregates, throughputs, and incomplete data. Accurate forecasts could help optimize resource management, anomaly detection, quality of service, and network configuration planning. Real network data will be used, and forecasting engines could provide outputs to assist other management tools. The supervisors are professors Cortez and Sousa from the University of Minho who have previous related research on internet traffic forecasting using neural networks.

Uploaded by

sidxxxnitc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
32 views1 page

PHD Proposal Title: Internet Traffic Forecasting Using Data Mining Techniques

This PhD proposal aims to develop advanced data mining techniques like neural networks and support vector machines to forecast internet traffic using limited information. The objectives are to forecast traffic for specific network topologies, aggregates, throughputs, and incomplete data. Accurate forecasts could help optimize resource management, anomaly detection, quality of service, and network configuration planning. Real network data will be used, and forecasting engines could provide outputs to assist other management tools. The supervisors are professors Cortez and Sousa from the University of Minho who have previous related research on internet traffic forecasting using neural networks.

Uploaded by

sidxxxnitc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

PhD Proposal Title: Internet Traffic Forecasting using Data Mining Techniques

Motivation and Objectives:



Internet traffic forecasting is an important issue for any medium/large network provider that has
received few attention from the Computer Networks research community. The aim of this PhD is to
study novel solutions for Internet traffic forecasting tasks, which may be used in distinct contexts
such as:
- Forecasting of Internet TCP/IP traffic, given a network topology with several nodes and links;
- Forecasting of specific Internet traffic aggregates (e.g. crucial flow aggregates, classes of service,
etc.)
- TCP/IP end-to-end throughputs predictions
- Forecasting scenarios where only incomplete information in available
- Improving Internet Traffic identification/classification tasks

These are challenging tasks, in particular when few information is available. The intention is to
develop advanced data mining techniques (e.g. Neural Networks or Support Vector Machines) that
can provide timely and accurate Internet traffic forecasts. Given this context, the results of this
research are expected to effectively contribute to enhance several Internet management related tasks
such as: general resource management procedures, improve anomaly detection efforts in the
Internet (e.g. spread of virus detection), increase the Quality of Service levels of Internet Providers,
assist intelligent optimization tools for configuration and planning of TCP/IP networks, among
many others.
Using the proposed forecasting techniques, prototypes of intelligent, user-configurable and versatile
Internet traffic forecasting engines could be implemented in order to provide useful outputs for
other Internet related management tools.

Additional notes: Although data collection often requires a huge effort/time, real-world data is/will
be available (e.g. UKERNA TCP/IP traffic - bytes/s of every link in the British academic network).

Supervisors:
- Paulo Cortez, [email protected], Department of Information Systems/Algoritmi R&D
Centre, Univ. Minho, Guimares, Portugal (http://www3.dsi.uminho.pt/pcortez)
- Pedro Sousa, [email protected], Department of Informatics/CCTC R&D, Univ. Minho,
Braga, Portugal (http://marco.uminho.pt/~pns)

Bibliography (previous work performed in this area):

- P. Cortez, M. Rio, M. Rocha and P. Sousa. Multiscale Internet Traffic Forecasting using Neural
Networks and Time Series Methods. In Expert Systems, Wiley-Blackwell, In press.
- P. Cortez, M. Rio, P. Sousa and M. Rocha. Forecasting Internet Traffic by Neural Networks
Under Univariate and Multivariate Strategies. In K. Madani (Ed.), Proceedings of 4th
International Workshop on Artificial Neural Networks and Intelligent Information Processing
(ANNIIP 2008), pp. 61-71, Funchal, Portugal, May, 2008. INSTICC Press
(http://hdl.handle.net/1822/8027).
- P. Cortez, M. Rio, P. Sousa and M. Rocha. Topology Aware Internet Traffic Forecasting using
- Neural Networks. In J. de S et al. (Eds.), Artificial Neural Networks - ICANN, 17th
International Conference, Lecture Notes in Computer Science 4669, pp. 445-454, Porto,
Portugal, September, 2007, Springer (http://hdl.handle.net/1822/7634).
- P. Cortez, M. Rio, M. Rocha and P. Sousa. Internet Traffic Forecasting using Neural Networks.
In Proceedings of the 2006 IEEE International Joint Conference on Neural Networks, pp. 4942-
4949, Vancouver, Canada, July, 2006 (http://hdl.handle.net/1822/6581).

You might also like