Sequential EM for Unsupervised Adaptive Gaussian Mixture Model Based Classifier

In this paper we present a sequential expectation maximization algorithm to adapt in an unsupervised manner a Gaussian mixture model for a classification problem. The goal is to adapt the Gaussian mixture model to cope with the non-stationarity in the data to classify and hence preserve the classification accuracy. Experimental results on synthetic data show that this method is able to learn the time-varying statistical features in data by adapting a Gaussian mixture model online. In order to control the adaptation method and to ensure the stability of the adapted model, we introduce an index to detect when the adaptation would fail.

[1]  Janez Demsar,et al.  Statistical Comparisons of Classifiers over Multiple Data Sets , 2006, J. Mach. Learn. Res..

[2]  Jdel.R. Millan,et al.  On the need for on-line learning in brain-computer interfaces , 2004, 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541).

[3]  Geoffrey E. Hinton,et al.  A View of the Em Algorithm that Justifies Incremental, Sparse, and other Variants , 1998, Learning in Graphical Models.

[4]  Shin Ishii,et al.  On-line EM Algorithm for the Normalized Gaussian Network , 2000, Neural Computation.

[5]  Pedro J. Moreno,et al.  A Study of Musical Instrument Classification Using Gaussian Mixture Models and Support Vector Machines , 1999 .

[6]  Christopher M. Bishop,et al.  Pattern Recognition and Machine Learning (Information Science and Statistics) , 2006 .

[7]  Atsushi Shimada,et al.  Dynamic Control of Adaptive Mixture-of-Gaussians Background Model , 2006, 2006 IEEE International Conference on Video and Signal Based Surveillance.

[8]  John Q. Gan,et al.  Unsupervised adaptive GMM for BCI , 2009, 2009 4th International IEEE/EMBS Conference on Neural Engineering.

[9]  Adrian E. Raftery,et al.  Incremental Model-Based Clustering for Large Datasets With Small Clusters , 2005 .

[10]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[11]  Carl E. Rasmussen,et al.  The Infinite Gaussian Mixture Model , 1999, NIPS.

[12]  Hsin-Min Wang,et al.  A Model-Selection-Based Self-Splitting Gaussian Mixture Learning with Application to Speaker Identification , 2004, EURASIP J. Adv. Signal Process..