Thomas, EM,Temko, A,Lightbody, G,Marnane, WP,Boylan, GB,

A Gaussian mixture model based statistical classification system for neonatal seizure detection

2009 IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING

2009

September

Validated

1

()

Neonatal Seizure Detection Linear Discriminant Analysis Principal Component Analysis Gaussian Mixture Models EEG

446

451

A neonatal seizure detection system is proposed based on a Gaussian mixture model classifier. Linear discriminant analysis and principal component analysis are compared for the task of feature vector preprocessing. A postprocessing scheme is developed from the probability of seizure estimate in order to improve the performance of the system. Results are reported on a dataset of 17 patients with a total duration of 267.9 hours, the average ROC area of the system is 95.6%.