[FieldTrip] multiclass svm

Bastiaansen, Marcel Bastiaansen4.M at nhtv.nl
Fri Nov 10 12:04:37 CET 2017


Dear Fieldtrippers,

As a relative novice to classification algorithms, I have been playing around with Fieldtrip's svm classifier (ft_statistics_crossvalidate). This seems to work pretty smoothly, but I have two questions related to that:

1.       In the given contingency table one can see the actual classification of trials. However, I assume the actual classification is based on a classification probability for each trial. Is there a way of getting a list of classification probabilities for all trials from this function? Else, a quick hint at where in the code these probabilities are being computed / stored would help me build this functionality

2.       I have EEG data that belong to 4 classes, but ft_statistics_crossvalidate does not afford multiclass approaches. Any tips or hints at what would be robust and well-validated approaches for multiclass models for EEG data would be much appreciated.

Best,
Marcel

***
Dr Marcel C.M. Bastiaansen
Senior lecturer and researcher in quantitative research methods
Academy for Leisure & Academy for Tourism
NHTV Breda University of Applied Sciences
Visiting adress:
Room C1.011, Academy for Leisure
Archimedesstraat 17,
4816 BA, Breda
Phone: +31 76 533 2869
Email: bastiaansen4.m at nhtv.nl<mailto:bastiaansen4.m at nhtv.nl>

And

Department of Cognitive Neuropsychology
Tilburg School of Social and Behavioral Sciences
Tilburg University
Visiting address:
Room S217, Simon building
Warandelaan 2
5000 LE Tilburg
Email: M.C.M.Bastiaansen at uvt.nl<mailto:M.C.M.Bastiaansen at uvt.nl>

publications<https://scholar.google.nl/citations?hl=en&view_op=list_works&gmla=AJsN-F5MJ8q_0925xM3GL1HwKMWkGTQGnfGN0OfdskoTXoB9cHMSkWFYZRW6Bp8iwksfnsTgAKylZkZ6NqoYRJj6skmxruDP8Q&user=u4iWSLoAAAAJ>
linked-in<https://www.linkedin.com/profile/view?id=AAMAAAHu0sABfdcxLjPvTgFyfucAZvQQcwfJXi0&trk=hp-identity-name>
***

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20171110/a5bb6e57/attachment-0001.html>


More information about the fieldtrip mailing list