2
90%
develop: 90%

Ran 20 Mar 2018 02:42AM UTC

Files 54

Run time 2s

Badge

Embed ▾

Committed 20 Mar 2018 02:38AM UTC coverage: 82.228%. Remained the same

Job # 972.2

Build Type

push

travis-ci

Committed by

bbengfort

Commit Message

Make sklearn classification dataset for Class Prediction Error Tests (#355)

In several places throughout the code, you will find that some test data is hardcoded.
This is the X/y data that is used by our models.  I originally used this data to test
my visualizer.  However, I soon found out that the data isn't suitable for all
applications.  In my case, using this data produced completely segmented clusters and this
isn't useful when you are trying to look at prediction error.

My solution is simple. I use the `sklearn.make_classification` function with a random state set
This allows for more complex data sets in which I am able to specify the number of classes
and how many are informative(plus more parameters).  I think the most important thing
is the setting the random state, this allows for completely reproducible images.

Run Details

1934 of 2352 relevant lines covered (82.23%)

0.82 hits per line

DistrictDataLabs / yellowbrick / 972 / 2
90%
develop: 90%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 972.2

DistrictDataLabs / yellowbrick / 972 / 2 90% develop: 90%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 972.2

DistrictDataLabs / yellowbrick / 972 / 2
90%
develop: 90%

README BADGES
x