These SVMs are needed to run 3 out of 4 implementations of neonatal seizure detection algorithms (https://github.com/ktapani/Neonatal_Seizure_Detection). The fullSVM_XXXs are trained on all data in https://zenodo.org/record/1280684#.Wxh3QkiFNaQ, and can be used to detect seizures in new EEG data. The svmXXXs are trained on cross validated data and can be used to reproduce results presented in: K. Tapani, S. Vanhatalo and N. Stevenson, Time-varying EEG correlations improve automated neonatal seizure detection, International Journal of Neural Systems. (accepted for publication)

2018