Include a summary of the training set (and test set) in the html file

[GoogleCodeExporter]

Include a summary of the training set (and test set) in the html file, as is 
done on the command line to STDOUT.

Just put what is printed to STDOUT in the html file as well, like:

Processing test set 'worm_part_classifier.fit'.
----------
Summary of 'worm_part_classifier.fit' (2036 samples total, 1 samples per image):
'Class label' number of samples.
'NewBackground' 197
'body'  828
'headALL'   739
'tails' 272
----------
Max balanced training: 197
samples per image=1, training images: 148, testing images 49

Original issue reported on code.google.com by christop...@nih.gov on 14 Mar 2011 at 5:52

[GoogleCodeExporter]

Looks like changes were made to the html report, to that very first table at 
the top that shows how many images were trained on and tested. Looks like it 
shows an aggregate instead of what the per split experiment set up was, as 
before.

With one split:

2036 Images in the dataset (samples per image=1)
    NewBackground   body    headALL     tails   total
Testing     49  49  49  49  196
Training    148     148     148     148     592

2036 Images in the dataset (samples per image=1)
    NewBackground   body    headALL     tails   total
Testing     245     245     245     245     980
Training    740     740     740     740     2960

Original comment by christop...@nih.gov on 14 Mar 2011 at 6:04

[GoogleCodeExporter]

Original comment by christop...@nih.gov on 14 Mar 2011 at 6:05

• Added labels: Priority-Medium
• Removed labels: Priority-Low

[GoogleCodeExporter]

So, change it back per-split?
Makes sense to move it to the bottom of the summary then, esp. considering the 
addition of the summary tables for the datasets at the top, which would have 
some overlapping info.

Original comment by i...@cathilya.org on 17 Mar 2011 at 5:40

• Changed state: Started

[GoogleCodeExporter]

Original comment by i...@cathilya.org on 18 Mar 2011 at 4:12

• Changed state: Fixed