<div dir="ltr">Hi all,<div><br></div><div><p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial">I’m having trouble
getting my head around something and I was wondering if you can give me a hand.<span></span></span></p><p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><br></span></p>

<p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial">I’m running a
classification with 4 possible categories, 10 runs. My data is balanced and I’m
using CSVM and a leave one out cross-validation.<span></span></span></p><p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><br></span></p>

<p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial">Just for fun, I wanted
to create a distribution of the possible performance if I randomized the labels
of the runs, so I was expecting a performance around 0.25, after 12,000 reps, I
got 0.200, I don’t get it, do you have any idea?<span></span></span></p><p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><br></span></p>

<p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial">This is part of the
code I used:<span></span></span></p><p class="MsoNormal"><span lang="EN-US" style="font-family:arial,sans-serif;background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"><br></span></p><p class="MsoNormal">clf = LinearCSVMC()</p><p class="MsoNormal">SensitivityBasedFeatureSelection(OneWayAnova(), FractionTailSelector(0.01, mode='select', tail='upper'))   </p><p class="MsoNormal">fclf = FeatureSelectionClassifier(clf, fsel)</p><p class="MsoNormal">cvte = CrossValidation(fclf, NFoldPartitioner(), errorfx=lambda p, t: np.mean(p == t), enable_ca=['stats'])</p><p class="MsoNormal">for k in range(0,rndReps):         </p><p class="MsoNormal"><span class="gmail-Apple-tab-span" style="white-space:pre">       </span>np.random.shuffle(fds.sa.targets)            </p><p class="MsoNormal"><span lang="EN-US" style="background-image:initial;background-position:initial;background-size:initial;background-repeat:initial;background-origin:initial;background-clip:initial"></span></p><p class="MsoNormal">        cv_results = cvte(fds)</p><p class="MsoNormal"><br></p><p class="MsoNormal">Thanks!</p><p class="MsoNormal"><br></p><p class="MsoNormal">Raul Hernandez</p></div></div>