Back to PLS Help
Correlation plot extraction
hsoderlund
Posted on 02/25/09 05:18:23
Number of posts: 0
Hi there,
I'm wondering if there's a way to extract the values of the correlation plot (including the confidence intervals!) after running a seed PLS, so that one can make the same figure in Excel.
Cheers,
Hedvig.
Replies:
Untitled Post
I'm Online
nlobaugh
Posted on 02/25/09 07:57:23
Number of posts: 229
see my earlier posting:
http://www.rotman-baycrest.on.ca/index.php?action=view_thread&id=518&module=bbmodule&src=%40random45c35fcb17881
nancy
Untitled Post
hsoderlund
Posted on 02/25/09 09:12:16
Number of posts: 0
That's great, it works! Thank you so much for your help!
Hedvig.
Untitled Post
hsoderlund
Posted on 02/25/09 09:12:30
Number of posts: 0
That's great. Thank you so much for your help!
Hedvig.
Untitled Post
hsoderlund
Posted on 02/26/09 08:03:43
Number of posts: 0
Hi again,
I ran this, but the numbers don't correspond to what I see in the correlation overview plot. From what I understand, if I look at LV5 I should look in column 5. And if I have 4 conditions, the data will be listed cond1 lag 1, cond1 lag2,.....,cond2 lag1...., right? I looked at both regular and adjusted CIs, but neither fit what's in the plot.
Perhaps it can be tricky to give advice without seeing my data, but I'd be grageful for any feedback/ideas at all.
Cheers,
Hedvig.
Untitled Post
I'm Online
jshen
Posted on 02/26/09 08:52:31
Number of posts: 291
The column corresponds to LV is correct, i.e. if you look at LV5 you should look in column 5.
However, the row corresponds to condition and behavior, i.e. cond1, behav1, cond1, behav2, ... condn,behav1, condn,behav2, ...
Untitled Post
hsoderlund
Posted on 02/26/09 08:56:03
Number of posts: 0
Yes, right, and here I did seed PLS, so it's lags instead of behaviors.
I'll keep searching....
Thanks Jimmy!
Untitled Post
I'm Online
jshen
Posted on 02/26/09 10:42:54
Number of posts: 291
The first lag was removed when you do Seed PLS. i.e. if you have 8 lags, 7 was used as Seed Behavior data.
If problem persistent, would you please let me know where you file sits? If it sits under:
/rri_disks/victoria/levine_lab/hedvig/new_data/pls_test
please tell me the result file name, so I can take a look into it for you.
Untitled Post
hsoderlund
Posted on 02/26/09 10:50:17
Number of posts: 0
Thank you Jimmy, that's really nice of you.
I just finally figured it out, and of course it was trivial: I thought the CIs would be indicated around the correlation (i.e. +/- the correlation value), but instead it's the value around 0. Silly...
Thanks again for your help,
Hedvig.
Login to reply to this topic.