confusion matrix labels in run classify lda exercise
120 views
Skip to first unread message
Maria Montchal
Jul 12, 2016, 4:52:24 PM7/12/16
to CoSMoMVPA
The matrix on the left is from the split half correlation exercise, and the one on the right is from the classify lda exercise. I understand that they are showing very different things, but I don't understand why the labels on the right (classify lda) are in such a strange order. It seems that "monkey" has been shifted to the bottom. Does anyone know why this is? is this correct...? the code I used to produce the confusion matrix on the right is copied and pasted directly from github, which I assume contains the current version. I'd appreciate any input!
Nick Oosterhof
Sep 26, 2016, 6:33:51 AM9/26/16
to CoSMoMVPA
Hello,
> On 12 Jul 2016, at 22:52, Maria Montchal <mmon...@gmail.com> wrote:
>
> The matrix on the left is from the split half correlation exercise, and the one on the right is from the classify lda exercise. I understand that they are showing very different things, but I don't understand why the labels on the right (classify lda) are in such a strange order. It seems that "monkey" has been shifted to the bottom. Does anyone know why this is? is this correct...? the code I used to produce the confusion matrix on the right is copied and pasted directly from github, which I assume contains the current version. I'd appreciate any input!
sorry for the late reply, this post initially escaped my attention.
I assume you are running the run_classify_lda.m script in the examples directory. When I run it the order is "correct", i.e. monkey, lemur, mallard, warbler, ladybug, lunamoth.
Do you use Matlab or Octave? Can you provide the output of cosmo_wtf please?
best,
Nick
> On 12 Jul 2016, at 22:52, Maria Montchal <mmon...@gmail.com> wrote:
>
> The matrix on the left is from the split half correlation exercise, and the one on the right is from the classify lda exercise. I understand that they are showing very different things, but I don't understand why the labels on the right (classify lda) are in such a strange order. It seems that "monkey" has been shifted to the bottom. Does anyone know why this is? is this correct...? the code I used to produce the confusion matrix on the right is copied and pasted directly from github, which I assume contains the current version. I'd appreciate any input!
I assume you are running the run_classify_lda.m script in the examples directory. When I run it the order is "correct", i.e. monkey, lemur, mallard, warbler, ladybug, lunamoth.
Do you use Matlab or Octave? Can you provide the output of cosmo_wtf please?
best,
Nick
Nick Oosterhof
Sep 29, 2016, 5:49:43 AM9/29/16
to Maria Montchal, CoSMoMVPA
Hello,
>>> On 12 Jul 2016, at 22:52, Maria Montchal <mmon...@gmail.com> wrote:
>>>
>>> The matrix on the left is from the split half correlation exercise, and the one on the right is from the classify lda exercise. I understand that they are showing very different things, but I don't understand why the labels on the right (classify lda) are in such a strange order. It seems that "monkey" has been shifted to the bottom.
>> On Monday, September 26, 2016 at 3:33:51 AM UTC-7, n.n.oosterhof wrote:
>> I assume you are running the run_classify_lda.m script in the examples directory. When I run it the order is "correct", i.e. monkey, lemur, mallard, warbler, ladybug, lunamoth.
>>
>> Do you use Matlab or Octave? Can you provide the output of cosmo_wtf please?
>
>>> On 12 Jul 2016, at 22:52, Maria Montchal <mmon...@gmail.com> wrote:
>>>
>>> The matrix on the left is from the split half correlation exercise, and the one on the right is from the classify lda exercise. I understand that they are showing very different things, but I don't understand why the labels on the right (classify lda) are in such a strange order. It seems that "monkey" has been shifted to the bottom.
>> On Monday, September 26, 2016 at 3:33:51 AM UTC-7, n.n.oosterhof wrote:
>> I assume you are running the run_classify_lda.m script in the examples directory. When I run it the order is "correct", i.e. monkey, lemur, mallard, warbler, ladybug, lunamoth.
>>
>> Do you use Matlab or Octave? Can you provide the output of cosmo_wtf please?
>
> On 28 Sep 2016, at 23:21, Maria Montchal <mmon...@gmail.com> wrote:
>
> yes, that is the script. I use Octave. cosmo_wtf output is attached. Thank you!
Thanks for the report, this was causes by not using 'XTick' prior to using 'XTickLabels', which gave different results in Matlab than in Octave. I've opened a pull request with updated examples that fix this issue [1]. This new code will most likely be merged in the master branch soon.
>
> yes, that is the script. I use Octave. cosmo_wtf output is attached. Thank you!
best,
Nick
[1] https://github.com/CoSMoMVPA/CoSMoMVPA/pull/92
Reply all
Reply to author
Forward
0 new messages