-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathNotes.txt
57 lines (49 loc) · 2.15 KB
/
Notes.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
To get rid of the "libpng warning: iCCP: known incorrect sRGB profile" use:
find . -name '*.png' -print -exec mogrify -strip {} \;
mogrify is from ImageMagick which can be installed on ubuntu/wsl using "sudo apt-get install imagemagick"
hdf5 library was a pain to compile. I ended up building it using Cmake on Windows to get the required H5pubconf.h file and running a build under VS2015 to generate the required H5lib_settings.c and H5Tinit.c. My guess is that this will need to be redone (probably using configure and make) for a Mac build.
Fonts in the icons are mostly in the Segoe family but converted to outlines for portability
Checking the LDA output
m_eigenvalues_sorted =
0.365689
0.0158325
-11.9832
-12.1707
-12.982
m_eigenvectors_sorted =
0.51367 0.0621404 0.193798 0.360615 0.017565
-0.627198 -0.654417 -0.517815 -0.739109 -0.331875
0.398345 -0.071636 -0.172426 0.472294 0.79278
-0.302021 0.741085 0.760623 -0.130064 -0.347357
-0.304747 -0.116353 -0.293313 0.289303 -0.374691
m_sd =
0.604722
0.125827
-nan(ind)
-nan(ind)
-nan(ind)
m_prop_of_var =
-0.00994954
-0.000430765
0.326034
0.331136
0.35321
m_cum_prop = -0.00994954
-0.0103803
0.315654
0.64679
1
Matlab
M
-0.0202607515370651 -0.0288087985606052 -0.00282811549716465 -0.0237392274288372 -0.00562659625499791
0.0527797691688551 0.0564076430497436 0.0351394170370974 0.0289870260541971 0.0592557691044090
0.0159777010277375 -0.0400040007359655 -0.0802511900967362 -0.0184142543489011 0.00648554524239449
-0.0737777805315444 0.0193212173337798 0.0351105614152083 0.0139618066913480 -0.0671023819286428
0.0266967674026525 -0.0301644134801936 0.0362051285768990 0.0140839140625211 0.0105354489875190
lambda
60.7844870209310
2.59627804521984
-0.111057645975554
-0.116056614413364
-0.117559585689291
Once these are normalised then the Matlab and the GalenQt eigenvectors are the same but oddly they are not sorted in the right order. Matlab 1 & 2 equates to GalenQt 2 & 1. And I'm pretty sure that Matlab is right since its image 1 contains the overtext. They are also reversed but that doesn't matter.