<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Dear Akiko,<div><br></div><div>you can use DML functions for multiclass classification: dml.one_against_one or dml.one_against_rest</div><div><br></div><div>one_against_rest <span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">builds one SVM per class, train</span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">ed to </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">distinguish the trials in a single class from the </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">trials in all remaining classes. </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">C</span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">lassification </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">of a test trial </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">is done according to the maximum </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">output among all SVMs.</span></div><div dir="ltr" style="font-size: 13.12px; font-family: sans-serif; left: 216px; top: 213.227px; transform: scale(0.89863, 1); transform-origin: 0% 0% 0px;" data-font-name="g_font_p0_1" data-canvas-width="3.2799986314773557"></div><div><br></div><div>one_against_one<span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; "> builds one SVM </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">for each pair of classes (a pairwise comparison). C</span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">lassification </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">of a test trial is done according to the </span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">maximum voting , where each SVM votes for one class</span><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">. </span></div><div><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; "><br></span></div><div><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">You can see for yourself which method is better for your research question/data. </span></div><div><span class="Apple-style-span" style="font-family: sans-serif; font-size: 13px; ">You call it like this: </span></div><div><font class="Apple-style-span" face="sans-serif"><span class="Apple-style-span" style="font-size: 13px;"></span></font></div><div><div>cfg.mva = dml.<wbr>one_against_rest('mva', {dml.standardizer() dml.svm()}) </div></div><div>or</div><div>cfg.mva = dml.<wbr>one_against_one('mva', {dml.standardizer() dml.svm()}) </div><div><br></div><div><br></div><div>best,</div><div>Irina</div><div><br></div><div><br></div><div><div><div><br></div></div></div><div> <br><div><div>On Apr 10, 2013, at 12:32 AM 4/10/13, Akiko Ikkai wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div dir="ltr">Hi,<div><br></div><div>I have a 3 conditions, A, B, and C in my MEG data. Multivariate classification using selected sensors between conditions A and B, B and C, and A and C work beautifully with chance = 50%. I'm now curious whether I could run a classification of 3 conditions with chance = 33%. </div>
<div><br></div><div>ft_freqstatistics could take more than 2 frequency datasets as inputs, but ft_statistics_crossvalidate using +dmlt/svm.m seems to require 2 datasets (because it's using a linear kernel?). <br><div>
<br></div><div style="">Does anyone have suggestions how I might go about it...?</div><div>Thanks in advance! Akiko<br clear="all"><div><br></div>-- <br><font><span style="font-family:arial,helvetica,sans-serif">Akiko Ikkai, Ph.D. <br>
Postdoctoral Fellow<br style="font-family:arial,helvetica,sans-serif"></span></font><font style="font-family:arial,helvetica,sans-serif" face="'PrimaSans BT,Verdana,sans-serif'">Department of
Psychological and Brain Sciences<br>Johns Hopkins University<br>Ames
Hall, 3400 N. Charles St.<br>Baltimore, MD 21218</font><br style="font-family:arial,helvetica,sans-serif"><br>
</div></div></div>
_______________________________________________<br>fieldtrip mailing list<br><a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</blockquote></div><br></div></body></html>