Online EM algorithm for mixture with application to internet traffic modeling

Authors:
Z. Liu;J. Almhana;V. Choulakian;R. McGorman
Affiliations:
Department of Computer Science, University of Moncton, Box 62, Moncton, New Brunswick, Canada E1A 3E9;Department of Computer Science, University of Moncton, Box 62, Moncton, New Brunswick, Canada E1A 3E9;Department of Computer Science, University of Moncton, Box 62, Moncton, New Brunswick, Canada E1A 3E9;Nortel Networks, 4001 E. Chapel Hill-Nelson Hwy, NC 27709, USA
Venue:
Computational Statistics & Data Analysis
Year:
2006

Citing 2
Cited 6

Self-Similar Network Traffic and Performance Evaluation

Self-Similar Network Traffic and Performance Evaluation
On convergence properties of the em algorithm for gaussian mixtures

Neural Computation

An online classification EM algorithm based on the mixture model

Statistics and Computing
A case-study on naïve labelling for the nearest mean and the linear discriminant classifiers

Pattern Recognition
Fast online graph clustering via Erdős-Rényi mixture

Pattern Recognition
Study of non-Markovian distributed primary radio activities on the opportunity time for secondary usage of spectrum

MILCOM'09 Proceedings of the 28th IEEE conference on Military communications
Prediction of web goodput using nonlinear autoregressive models

IEA/AIE'10 Proceedings of the 23rd international conference on Industrial engineering and other applications of applied intelligent systems - Volume Part II
Self-similarity of the simulated internet traffic from the source with clustered fragmented periodic output

Proceedings of the 2013 Summer Computer Simulation Conference

Quantified Score

Hi-index	0.03

Visualization

Abstract

Since histograms of many real network traces show strong evidence of mixture, this paper uses mixture distributions to model Internet traffic and applies the EM algorithm to fit the models. Making use of the fact that at each iteration of the EM algorithm the parameter increment has a positive projection on the gradient of the likelihood function, this paper proposes an online EM algorithm to fit the models and the Bayesian Information Criterion is applied to select the best model. Experimental results on real traces are provided to illustrate the efficiency of the proposed algorithm.