2,684
edits
Line 96: | Line 96: | ||
</pre> | </pre> | ||
The output is: <vector #> <x> <y> <length> <angle> for each vector, in this 2-dimensional case; equivalently for higher dimensions. | The output is: <vector #> <x> <y> <length> <angle> for each vector, in this 2-dimensional case; equivalently for higher dimensions. | ||
The Python code produces the following output: | |||
<pre> | |||
=== | |||
Correlation matrix parsed from infile: | |||
[[ nan 0.017 0.0222 0.0233 0.0226] | |||
[0.017 nan 0.7026 0.7287 0.6241] | |||
[0.0222 0.7026 nan 0.9131 0.8049] | |||
[0.0233 0.7287 0.9131 nan 0.8432] | |||
[0.0226 0.6241 0.8049 0.8432 nan]] | |||
=== | |||
Correction factor for 2nd and higher eigenvalue(s): | |||
0.8000 | |||
=== | |||
Interpretation of correlation matrix as dot product matrix: | |||
--- | |||
all h_i by iterative approach: | |||
initial values: | |||
[0.1459 0.7198 0.7815 0.7919 0.7574] | |||
refinement by iteration: | |||
#13: [0.0242 0.7414 0.9389 0.9798 0.8551] | |||
=== | |||
Uncorrected eigenvalue(s): | |||
2 used: | |||
[3.1228 0.0126] | |||
3 unused: | |||
[ 0.0045 -0.0004 -0.0167] | |||
--- | |||
Corrected eigenvalue(s): | |||
2 used: | |||
[3.1228 0.0158] | |||
iter RMS max_chg rms_chg | |||
0 0.00345 - - | |||
1 0.00241 -0.04680 0.00403 | |||
2 0.00127 0.01783 0.00275 | |||
3 0.00057 0.01297 0.00162 | |||
4 0.00029 0.00570 0.00073 | |||
5 0.00023 0.00182 0.00026 | |||
6 0.00023 -0.00047 0.00008 | |||
7 0.00022 -0.00042 0.00004 | |||
8 0.00022 -0.00042 0.00004 | |||
9 0.00022 -0.00045 0.00005 | |||
10 0.00022 -0.00045 0.00005 | |||
11 0.00022 -0.00045 0.00005 | |||
12 0.00021 -0.00044 0.00005 | |||
13 0.00021 -0.00043 0.00005 | |||
14 0.00021 -0.00043 0.00005 | |||
15 0.00021 -0.00042 0.00004 | |||
16 0.00021 -0.00042 0.00004 | |||
17 0.00020 -0.00042 0.00004 | |||
18 0.00020 -0.00042 0.00004 | |||
19 0.00020 -0.00042 0.00004 | |||
20 0.00020 -0.00042 0.00004 | |||
1 0.0241 -0.0098 | |||
2 0.7484 0.1425 | |||
3 0.9359 0.0150 | |||
4 0.9758 -0.0112 | |||
5 0.8624 -0.1501 | |||
=== | |||
Finished outputting 2-dimensional representative vectors! =) | |||
</pre> |