[MRG+2] Rename `MICEImputer` to `ChainedImputer` by sergeyf · Pull Request #11314 · scikit-learn/scikit-learn

sergeyf · 2018-06-18T16:34:34Z

Addresses a point raised in #8478 and #11259. The MICEImputer isn't really a MICE imputer because we're only keeping a single averaged imputation by default. This is a purely cosmetic PR so that end-users don't get as confused.

CC @glemaitre @jnothman @RianneSchouten

jnothman

LGTM

sergeyf · 2018-06-20T00:02:47Z

Getting some urllib.error.HTTPError: HTTP Error 500: Internal Server Error =(. Will wait a few hours and try again.

qinhanmin2014 · 2018-06-20T03:07:33Z

@sergeyf Master is failing, please merge master in after #11318 is merged.

qinhanmin2014 · 2018-06-20T03:11:51Z

@sergeyf Any reference for the new name?
Also, I still find things like A more sophisticated approach is to use the :class:`ChainedImputer` class, which implements the the technique of Multivariate Imputation by Chained Equations. MICE..., which makes me confused.

sergeyf · 2018-06-20T04:42:12Z

@qinhanmin2014 thanks re: master, I'll do that.

Re: name. So the point here is that we're not actually doing MICE, because the traditional way of applying the MICE technique is to get multiple imputations, and this class only returns a single one. But the full name is multivariate imputation by chained equations (MICE), hence the named ChainedImputer. I guess I can put that language in directly.

Fix test failures

jnothman · 2018-06-20T05:37:53Z

Sorry, I broke master.

jnothman · 2018-06-20T05:39:16Z

Part of the confusion is that:

the paper says MICE = Multiple Imputation by Chained Equations
the R implementation says MICE = Multivariate Imputation by Chained Equations

sergeyf · 2018-06-20T13:51:26Z

@jnothman Which paper? Stef's paper that I reference is this one: "mice: Multivariate Imputation by Chained Equations in R" https://www.jstatsoft.org/article/view/v045i03

There's also this paper, but it's not referenced in this PR: "Multiple Imputation by Chained Equations: What is it and how does it work?" https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3074241/

@qinhanmin2014 will you please take a look at my changes to impute.rst and let me know if it's sufficiently clear now.

jnothman · 2018-06-20T14:09:52Z

You're right. I've just been confused by people implying that the M must mean Multiple.... because of course what we're doing here is Multivariate Imputation by Chaining.... just not Multiple.

sergeyf · 2018-06-20T14:29:08Z

I have confused that at least 5 times while writing the original PR =)

ogrisel · 2018-06-20T14:39:30Z

doc/modules/impute.rst

+A more sophisticated approach is to use the :class:`ChainedImputer` class, which
+implements MICE: Multivariate Imputation by Chained Equations. MICE is usually used
+to generate multiple imputations, but :class:`ChainedImputer` generates a single
+(averaged) imputation, which is why it is not named `MICEImputer`. MICE models each


I think we need to link those terms to a glossary entry that explains:

univariate vs multivariate imputation

single vs multiple imputation (increasing the number of samples in the resulting transformed dataset).

Good idea. There is no entry for imputation. I'll add one and include these points.

ogrisel

Besides the above comment on the missing glossary entry, +1 for merge.

sergeyf · 2018-06-20T15:17:49Z

@ogrisel I added a glossary entry and rewrote the entries in impute.rst. Please take a look and confirm that it satisfies your request.

RianneSchouten · 2018-06-20T16:39:41Z

Yeah, mice is both multivariate and multiple. multivariate: it loops over the features and imputes all incomplete features multiple: if you impute every incomplete value m times. ChainedImputer in its current form is chained and multivariate (and not yet multiple). <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> Virusvrij. www.avast.com <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> 2018-06-20 17:19 GMT+02:00 Sergey Feldman <notifications@github.com>:

…

@ogrisel <https://github.com/ogrisel> I added a glossary entry and rewrote the entries in impute.rst. Please take a look and confirm that it satisfies your request. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#11314 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AVQqeyAS6OX5LzBwe2JtzlCeq2jly0RCks5t-mfqgaJpZM4UsF-9> .

ogrisel

The glossary entry for imputation is significantly longer than the average entry.

Maybe we should keep a short entry in the glossary and move the discussing on univariate vs multivariate vs single vs multiple imputations in a dedicated section at the end of the missing value imputation chapter of the narrative documentation.

Alternatively, we could split the entries: univariate imputation, multivariate imputation, multiple imputations and add cross-links. See for instance the entries on "inductive" vs "transductive".

ogrisel · 2018-06-21T09:38:08Z

doc/glossary.rst

+    impute
+    imputation
+        Most machine learning algorithms require that their inputs have no
+        :term:`missing values`, and will not work if this requirement is


Please add a backlink to the :term:imputation entry from the :term:missing values entry.

ogrisel · 2018-06-21T09:49:11Z

doc/glossary.rst

+        The above practice is called multiple imputation. The
+        :class:`impute.ChainedImputer` class can be used for multiple
+        imputations by applying it to the same dataset with different random
+        seeds.


I would further refine:

... imputations by applying it repeatedly to the same dataset with different random seeds.
Note that a call to the transform method of a :term:transformer is not allowed to change the number of samples. Therefore multiple imputations cannot be achieved by a single call to transform.

sergeyf · 2018-06-21T15:03:17Z

@ogrisel Done! I moved most of the discussion to impute.rst

sergeyf · 2018-06-21T15:55:03Z

Hmm tests are failing because of StandardScaler?

=================================== FAILURES ===================================
 test_non_meta_estimators[StandardScaler-StandardScaler-check_estimators_nan_inf] 
name = 'StandardScaler'
Estimator = <class 'sklearn.preprocessing.data.StandardScaler'>
check = <function check_estimators_nan_inf at 0x7fe165faca60>

glemaitre · 2018-06-21T16:04:42Z

sklearn/utils/estimator_checks.py


-ALLOW_NAN = ['Imputer', 'SimpleImputer', 'MICEImputer',
-             'MinMaxScaler', 'StandardScaler', 'QuantileTransformer']
+ALLOW_NAN = ['Imputer', 'SimpleImputer', 'ChainedImputer',


Add the StandardScaler back here to remove the error.

Oh dang that was my bad on the merging. Thanks!

ogrisel · 2018-06-22T08:01:28Z

Thank you very much @sergeyf!

This reverts commit 93382cc.

initial commit

c35fd63

This was referenced Jun 18, 2018

MICE/Multiple Imputation branch #11259

Closed

[MRG+2] Basic version of MICE Imputation #8478

Merged

sergeyf added 2 commits June 18, 2018 12:33

fixing failure due to spacing in impute.rst

ea4e83f

fixing failure due to spacing in impute.rst

3c26b6b

sergeyf changed the title ~~Rename MICEImputer to ChainedImputer~~ [MRG] Rename MICEImputer to ChainedImputer Jun 18, 2018

jnothman approved these changes Jun 19, 2018

View reviewed changes

sergeyf changed the title ~~[MRG] Rename MICEImputer to ChainedImputer~~ [MRG+1] Rename MICEImputer to ChainedImputer Jun 19, 2018

Improving language in impute.rst

760c179

sergeyf closed this Jun 19, 2018

sergeyf reopened this Jun 19, 2018

sergeyf closed this Jun 20, 2018

sergeyf reopened this Jun 20, 2018

sergeyf closed this Jun 20, 2018

sergeyf reopened this Jun 20, 2018

sergeyf closed this Jun 20, 2018

sergeyf reopened this Jun 20, 2018

sergeyf added 2 commits June 19, 2018 21:50

updating impute,rst some more

f0b717d

M:class:ChainedImputer merge branch 'master' into mice_rename

ca1b462

Fix test failures

ogrisel reviewed Jun 20, 2018

View reviewed changes

ogrisel approved these changes Jun 20, 2018

View reviewed changes

ogrisel added this to the 0.20 milestone Jun 20, 2018

adding glossary and cleaning up impute.rst

7e8ed30

sergeyf changed the title ~~[MRG+1] Rename MICEImputer to ChainedImputer~~ [MRG+2] Rename MICEImputer to ChainedImputer Jun 20, 2018

ogrisel reviewed Jun 21, 2018

View reviewed changes

sergeyf and others added 3 commits June 21, 2018 07:56

Merge branch 'master' into mice_rename

ee3c81a

updating docs again

018edae

merge conflict resolved

6d4ce49

Merge branch 'master' into mice_rename

8245041

glemaitre reviewed Jun 21, 2018

View reviewed changes

sergeyf and others added 3 commits June 21, 2018 09:15

undoing foolishness

9a44859

fixing impute.rst again

22a3393

Merge branch 'master' into mice_rename

9dbba54

ogrisel merged commit 93382cc into scikit-learn:master Jun 22, 2018

sergeyf deleted the mice_rename branch June 22, 2018 15:48

RianneSchouten mentioned this pull request Jun 27, 2018

[WIP] Multiple Imputation: Example with IterativeImputer #11370

Closed

jorisvandenbossche added a commit to jorisvandenbossche/scikit-learn that referenced this pull request Jul 17, 2018

Revert "Rename MICEImputer to ChainedImputer (scikit-learn#11314)"

486c079

This reverts commit 93382cc.

Uh oh!

Conversation

sergeyf commented Jun 18, 2018

Uh oh!

jnothman left a comment

Choose a reason for hiding this comment

Uh oh!

sergeyf commented Jun 20, 2018

Uh oh!

qinhanmin2014 commented Jun 20, 2018

Uh oh!

qinhanmin2014 commented Jun 20, 2018

Uh oh!

sergeyf commented Jun 20, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jnothman commented Jun 20, 2018

Uh oh!

jnothman commented Jun 20, 2018

Uh oh!

sergeyf commented Jun 20, 2018

Uh oh!

jnothman commented Jun 20, 2018

Uh oh!

sergeyf commented Jun 20, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ogrisel left a comment

Choose a reason for hiding this comment

Uh oh!

sergeyf commented Jun 20, 2018

Uh oh!

RianneSchouten commented Jun 20, 2018 via email

Uh oh!

ogrisel left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sergeyf commented Jun 21, 2018

Uh oh!

sergeyf commented Jun 21, 2018

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ogrisel commented Jun 22, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

sergeyf commented Jun 20, 2018 •

edited

Loading

ogrisel left a comment •

edited

Loading