4
55%
master: 78%

Ran 28 Sep 2017 10:58PM UTC

Files 88

Run time 4s

Badge

Embed ▾

Committed 28 Sep 2017 10:44PM UTC coverage: 78.397% (+0.7%) from 77.648%

Job # TOX_ENV=coveralls

Build Type

push

travis-ci

Committed by

qimingj

Commit Message

ML Workbench: If top-n is 0, instead of outputing "predicted, predicted_2,...probability, probability_2,...", output "predicted, class1, class2,...".

For example, if top-n is 0, before the output was:

predicted: daisy
predicted_2: roses
predicted_3: sunflower
probability: 0.9
probability_2: 0.1
probability_3: 0.1

With this change it will be:
predicted: daisy
daisy: 0.9
roses: 0.1
sunflower: 0.1

The reason for the change is that if all classes are included, top-n results are not that useful. But in some cases, it is convenient to have a column of a certain class's probability. For example, plotting ROC curve requires prob of a certain class.

Run Details

4772 of 6087 relevant lines covered (78.4%)

0.78 hits per line

googledatalab / pydatalab / 2354 / 4
55%
master: 78%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 2354.4 (TOX_ENV=coveralls)

googledatalab / pydatalab / 2354 / 4 55% master: 78%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 2354.4 (TOX_ENV=coveralls)

googledatalab / pydatalab / 2354 / 4
55%
master: 78%

README BADGES
x