Hi everyone, I'm doing research about BCI , until now, I used the OpenVibe Software v1.2.2 but the yesterday I downloaded the new version 2.0.1, my surprise have been when I try my escenarios that I did with openvibe v1.2.2 with the new version 2.0.1, the results are distincts but a lot distincts.
I don't knoew why.
I attach the two escenarios (v1.2.2 and v2.0.1 and the results), the data are the same in the two versions.
Escenario v2.0.1:
Escenario v1.2.2:
Results:
What could be the problem?
Differences between openvibe v.1.2.2 and v2.0.1
-
- Posts: 775
- Joined: Tue Dec 04, 2012 3:53 pm
- Location: INRIA Rennes, FRANCE
Re: Differences between openvibe v.1.2.2 and v2.0.1
Hi, could it be the same reason as is discussion in this thread?
viewtopic.php?f=17&t=9845
Can you try the suggestion there and let us know if it changes the situation?
Also my standard addition to any post concerning cross-validation result interpretation,
http://openvibe.inria.fr/tutorial-how-t ... te-better/
Cheers,
Jussi
viewtopic.php?f=17&t=9845
Can you try the suggestion there and let us know if it changes the situation?
Also my standard addition to any post concerning cross-validation result interpretation,
http://openvibe.inria.fr/tutorial-how-t ... te-better/
Cheers,
Jussi
Re: Differences between openvibe v.1.2.2 and v2.0.1
Hi, thanks for you response.
But I saw that exist difference in temporal filter between v1.2.2 and v2.0.1, I dont know why, but I think that this factor affect the accuracy.
But I saw that exist difference in temporal filter between v1.2.2 and v2.0.1, I dont know why, but I think that this factor affect the accuracy.
-
- Posts: 775
- Joined: Tue Dec 04, 2012 3:53 pm
- Location: INRIA Rennes, FRANCE
Re: Differences between openvibe v.1.2.2 and v2.0.1
If thats the case, you can use the "Temporal Filter (INSERM contrib)" which is the old temporal filter. In any case please let us know if the problem is fixed by these changes or not.
If there's anything wrong with the classifier trainer or the temporal filter in 2.0.1, we of course need to fix them. But first we need to pinpoint the problem.
After you have tried these suggestions and if they don't work, perhaps you could send us an example data file and the associated scenarios?
Also note that you *really* need to update the boxes that say 'update' or they might be working wrong.
Best,
Jussi
If there's anything wrong with the classifier trainer or the temporal filter in 2.0.1, we of course need to fix them. But first we need to pinpoint the problem.
After you have tried these suggestions and if they don't work, perhaps you could send us an example data file and the associated scenarios?
Also note that you *really* need to update the boxes that say 'update' or they might be working wrong.
Best,
Jussi