DistrictDataLabs / yellowbrick / 429 / 2
90%
develop: 90%

DEFAULT BRANCH: develop
Ran
Files 77
Run time 5s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 80.457% (+0.8%) from 79.614%
429.2

push

travis-ci-com

lwgray 
Finishing up CVscores visualizer (#559)

This PR aims to complete the tasks in #464 to finalize @pdamodaran's new (and most useful!) CVScores visualizer. Here's what's done:

* Modified the legend that labels the average cross-validation score dotted line with the numeric value of self.cv_scores_mean_

* Removed the ylim constraint so that CVScores plots will be more easily comparable across different models.

* Renames the module from cv to cross_validation for readability and accessibility to newcomers.

* Some additional cleanup in the docstrings for nice rendering in the documentation.

* Updates the docs with the new names and new plots with unadjusted ylim and updated legend.

* Adds some preliminary test coverage, all of which pass on my os ;)

Below are all commit titles
--------------------------
* encountering some issues when trying to use the size parameter to set the size of the figure on instantiation of the visualizer

* figured out the sizing issue

* coef shape for logistic regression needs to be flattened

* sklearn logistic regression causing issue for feature importances due to difference coef_ shape

* adds mean score to CVScores legend, removes ylim modification to make plots more comparable, updates from cv to cross_validation for readability

* in CVScores docs, updates from cv to cross_validation for readability

* updates cross validation docs from cv to cross_validation, shows new images with mean score in legend and no adjustment of ylim

* adds preliminary tests for CVScores and new baseline images

* fix test docstring

* increasing the tolerance for the regression test to try to resolve the images not close error

* ups the tolerance for the regression image test to 33

* adds tests to ensure CVScores correctly passes cv param to internal sklearn classifiers or regressors that have a CV param

* increases the tolerance for the regression image closeness to 36

2709 of 3367 relevant lines covered (80.46%)

0.8 hits per line

Source Files on job 429.2