DOC replace boston plot_outlier_detection#16687
DOC replace boston plot_outlier_detection#16687glemaitre merged 4 commits intoscikit-learn:masterfrom
Conversation
adrinjalali
left a comment
There was a problem hiding this comment.
Looks pretty good. Thanks. Could you maybe increase the width of the plots so that the legend fits in the plot maybe?
Also, while you're at it, it'd be really nice if you could divide the example into sections for the two examples the same way it's done in this example for instance.
|
I like the new data. @lucyleeow since you know perfectly |
|
Have separated each example and plot so that they are together and the legend fits in the plot now. Thanks @glemaitre and @adrinjalali |
|
ping @glemaitre |
adrinjalali
left a comment
There was a problem hiding this comment.
Thanks @lucyleeow . LGTM.
|
Thanks @lucyleeow |
Towards #16155
What does this implement/fix? Explain your changes.
Replaces data in
plot_outlier_detection_housing.pyfrom boston to wine recognition.Previous:
New:
The first example is slightly different (as the observations confounded into one was quite unique to the boston data set) but still shows a main cluster and deviation of non robust empirical covariance estimate. I assumed that 'confounding' was referring to data problems resulting in many observations being the same so changed the explanatory text. In the new example it is more a main 'primary' cluster with outlying points, possibly another small cluster.
ping @ogrisel