What steps will reproduce the problem?
1. Run a train [seems OK: .sig files created, fit file populated]
2. Run a classify [seems OK: .sig files created]
3. See classify result, which is definitely not OK!
What is the expected output? What do you see instead?
Part of the output:
[...]
classifyXIX/9867-04ep.tif 2.25e-19 1.000 0.000 * beteg 0.000
classifyXIX/9872-04_I_ep.tif 2.25e-19 1.000 0.000 * beteg 0.000
classifyXIX/9872-04_I_ep_2.tif 2.25e-19 1.000 0.000 * beteg 0.000
Average accuracy (1 splits): -nan
Confusion Matrix (sum of all splits)
beteg ep Total Tested Per-Class Accuracy
beteg 0 0 0 -nan
ep 0 0 0 -nan
Seems like no classifycation has been done.
Norm fact is always: 2.25e-19 probabilities are 1.000 and 0.000, which is I
assume needless to say is not the expected result.
What version of the product are you using? On what operating system?
1.50.727, on Linux debian 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1 x86_64
GNU/Linux, also occured on Linux Sabayon 3.8.0-sabayon #1 SMP Fri May 24
17:20:54 UTC 2013 x86_64 Intel(R) Core(TM) i3 CPU M 370 @ 2.40GHz GenuineIntel
GNU/Linux
Please provide any additional information below.
.SIG calculations may be good (I cannot check though), but the classification
decision calculation may be broken.
This problem not occur on 1.30.227 (we use that version). This does not mean
this thing has been broken since then, we did not check the middle versions.
Original issue reported on code.google.com by pinterg...@gmail.com on 23 Jul 2013 at 8:32
Original issue reported on code.google.com by
pinterg...@gmail.comon 23 Jul 2013 at 8:32