help needed with P300Classifier

Forum for discussion on different signal processing algorithms
Locked
Ray Ma
Posts: 14
Joined: 15 Feb 2010, 16:39

help needed with P300Classifier

Post by Ray Ma » 15 Apr 2010, 18:00

Hi,

I can generate SWLDA weights using fake data generated by the signal generator, but always get the following error, "Signal must have rows > cols. Linear classifier is valid only for overdetermined systems", when I tried to process the data collected from our IOTech ADC.

I'll post the link to the data and the parameter here:
http://dl.dropbox.com/u/1783209/rm_onlineS330R05.dat
http://dl.dropbox.com/u/1783209/P3OnlineSpell.prm

Can somebody help me take a look?

When I tried to import the fake as well as the real data into matlab, I found that the fake signal amplitude are int16 numbers of large values, whereas the real data has amplitude single precision floating point numbers of very small values. The reason for this might be the SourceGain was set differently for them. I don't know whether that matters or not...

Thanks very much!

Ray

Ray Ma
Posts: 14
Joined: 15 Feb 2010, 16:39

Re: help needed with P300Classifier

Post by Ray Ma » 22 Apr 2010, 01:19

I have modified P300GUI, the matlab counterpart of P300Classifier to accommodate my experimental paradigm. Seems like my P3Speller is not messing up the StimulusCode or StimulusType, and P300GUI is generating MUD files. I still need to test its performance experimentally.
Ray Ma wrote:Hi,

I can generate SWLDA weights using fake data generated by the signal generator, but always get the following error, "Signal must have rows > cols. Linear classifier is valid only for overdetermined systems", when I tried to process the data collected from our IOTech ADC.

I'll post the link to the data and the parameter here:
http://dl.dropbox.com/u/1783209/rm_onlineS330R05.dat
http://dl.dropbox.com/u/1783209/P3OnlineSpell.prm

Can somebody help me take a look?

When I tried to import the fake as well as the real data into matlab, I found that the fake signal amplitude are int16 numbers of large values, whereas the real data has amplitude single precision floating point numbers of very small values. The reason for this might be the SourceGain was set differently for them. I don't know whether that matters or not...

Thanks very much!

Ray

Locked

Who is online

Users browsing this forum: No registered users and 21 guests