Change colors of confusion matrix to shades of blue and change the axis order to match scikit-learn's#426
Conversation
Codecov Report
@@ Coverage Diff @@
## master #426 +/- ##
=======================================
Coverage 97.47% 97.47%
=======================================
Files 105 105
Lines 3287 3290 +3
=======================================
+ Hits 3204 3207 +3
Misses 83 83 Continue to review full report at Codecov.
|
kmax12
left a comment
kmax12
left a comment
There was a problem hiding this comment.
this looks good. one other change would be to change the axis order to match scikit learn as well. so the correct predicts are the diagonal that go top left to bottom right.
angela97lin
changed the title
|
@kmax12 done! I updated the image I attached with the new axis order :) |
|
@kmax12 okay, I think I actually updated the plot to match sklearn's now (see attached photo) 👍 |
| labels = conf_mat.columns | ||
|
|
||
| # reverse columns in confusion matrix to change axis order to match sklearn's | ||
| conf_mat = conf_mat.iloc[:, ::-1] |
There was a problem hiding this comment.
does it make sense to do this here or in get_confusion_matrix_data function? ideally that function returns the data as closely as possible to what makes it easy to create the graph
There was a problem hiding this comment.
Hmmm, I agree that given what we previously talked about (generate_confusion_matrix should simply take in the data and graph it) that makes sense. I've updated the PR to move the reversal to get_confusion_matrix_data though I wonder if the confusion matrix data feels less intuitive to read now 🤔 What do you think?
2 participants
Closes #425.
