Affiliation:
1. Angers University, Angers, France
2. Lebanese University, Saida, Lebanon
Abstract
The aim of this paper is to find the best intelligent model that allows predicting the future of premature newborns according to their electroencephalogram (EEG). EEG is a signal that measures the electrical activity of the brain. In this paper, the authors used a dataset of 397 EEG records detected at birth of premature newborns and their classification by doctors two years later: normal, sick or risky. They executed machine learning on this dataset using several intelligent models such as multiple linear regression, linear discriminant analysis, artificial neural network and decision tree. They used 14 parameters concerning characteristics extracted from EEG records that affect the prognosis of the newborn. Then, they presented a complete comparative study between these models in order to find who gives best results. Finally, they found that decision tree gave best result with performance of 100% for sick records, 76.9% for risky and 69.1% for normal ones.
Reference21 articles.
1. Lieshout, H. V., Jacobs, J. W. F. M., Rotteveel, J. J., Geven, W., & Hof, M. V. T. (1995). The prognostic value of the EEG in asphyxiated newborns. ActaNeurolScand,91(3), 203–207.
2. Benders, M. J., Palmu, K., Menache, C., Borradori-Tolsa, C., Lazeyras, F., Sizonenko, S., ... & Hüppi, P. S. (2014). Early Brain Activity Relates to Subsequent Brain Growth in Premature Infants.
3. Chauvet, P., & Nguyen, S. (2013). EEGDiag, une application d’analyse de l’EEG pour la plateforme de télésanté BBEEG. 4ièmes Journées Démonstrateurs.
4. Automatic Burst Detection based on Line Length in the Premature EEG;N.Koolen;International Conference on Bio-inspired Systems and Signal Processing,2013).