Code

[MRG] Add quantile regression#9978

Code

Merged

glemaitre merged 147 commits intoscikit-learn:mainscikit-learn/scikit-learn:mainfrom

avidale:quantile-regressionavidale/scikit-learn:quantile-regressionCopy head branch name to clipboard

May 25, 2021

Commits

Commits on Oct 4, 2017

• Added basic Quantile regression

  

  David Dale

  committed
  7988049

  Copy full SHA for 7988049

• Added L1 penalty to loss function

  

  David Dale

  committed
  fd04aed

  Copy full SHA for fd04aed

Commits on Oct 16, 2017

• Add tests and plot example for Quantile Regression

  

  David Dale

  committed
  6284c83

  Copy full SHA for 6284c83

Commits on Oct 17, 2017

• Enable gtol and maxiter

  

  David Dale

  committed
  8ed4899

  Copy full SHA for 8ed4899

• light refactor

  

  David Dale

  committed
  285b5d7

  Copy full SHA for 285b5d7

• Code for approximate smooth quantile loss

  

  David Dale

  committed
  bb5ce51

  Copy full SHA for bb5ce51

• Fix error in the smooth version

  

  David Dale

  committed
  fb07f0e

  Copy full SHA for fb07f0e

Commits on Oct 18, 2017

• Solve a sequence of smooth problems instead of one non-smooth.

  

  David Dale

  committed
  95492fe

  Copy full SHA for 95492fe

• Tuned convergence

  

  David Dale

  committed
  bdb48ac

  Copy full SHA for bdb48ac

• Rename approximation threshold, write docstrings

  

  David Dale

  committed
  8ee950d

  Copy full SHA for 8ee950d

• Enforce zero coefficients

  

  David Dale

  committed
  00a72fb

  Copy full SHA for 00a72fb

• Fix zero enforcement

  

  David Dale

  committed
  4a25fad

  Copy full SHA for 4a25fad

Commits on Oct 22, 2017

• Add description in user guide

  

  David Dale

  committed
  b3ccff6

  Copy full SHA for b3ccff6

• pep8 line widths

  

  David Dale

  committed
  cb87e63

  Copy full SHA for cb87e63

• Merge remote-tracking branch 'upstream/master' into quantile-regression

  

  David Dale

  committed
  0ea2317

  Copy full SHA for 0ea2317

• Mention GradientBoostingRegressor in the docs.

  

  David Dale

  committed
  b567c82

  Copy full SHA for b567c82

• Mention robustness in doc.

  

  David Dale

  committed
  0686e66

  Copy full SHA for 0686e66

Commits on Mar 18, 2018

• Remove magic constants for quantile regression

  

  David Dale

  committed
  d2a59d8

  Copy full SHA for d2a59d8

• Small edits to .rst on quantile regression

  

  David Dale

  committed
  4dc4eb2

  Copy full SHA for 4dc4eb2

• Improve formatting of quantile regression tests

  

  David Dale

  committed
  eca1165

  Copy full SHA for eca1165

• Change the solver for quantile regression

  avidale

  committed
  2524c2b

  Copy full SHA for 2524c2b

• small refactor of tests of the quantile regression

  

  David Dale

  committed
  04e6165

  Copy full SHA for 04e6165

• Merge branch 'quantile-regression' of https://github.com/avidale/scikit-learn into quantile-regression

  

  David Dale

  committed
  4b715d3

  Copy full SHA for 4b715d3

• Enable normalization of inputs

  

  David Dale

  committed
  52e1aa1

  Copy full SHA for 52e1aa1

Commits on Mar 19, 2018

• Change convergence warning and max_iter in some tests

  

  David Dale

  committed
  8fa745b

  Copy full SHA for 8fa745b

• A small refactor to the quantile regression example and docs

  

  David Dale

  committed
  90bc013

  Copy full SHA for 90bc013

• Comparison of QuantileRegressor and OLS

  

  David Dale

  committed
  06c24c3

  Copy full SHA for 06c24c3

• update test_quantile_warm_start

  

  David Dale

  committed
  9b0d42c

  Copy full SHA for 9b0d42c

• Improved the warm_start example, added a toy example for quantile regression

  

  David Dale

  committed
  5e453aa

  Copy full SHA for 5e453aa

• fig the toy test case for quantile regression

  

  David Dale

  committed
  463c633

  Copy full SHA for 463c633

Commits on Oct 7, 2018

• Merge branch 'master' into quantile-regression

  

  ddale

  committed
  80c8182

  Copy full SHA for 80c8182

• respond to comments and fix some style

  

  ddale

  committed
  50a5105

  Copy full SHA for 50a5105

• fix the gradient calculation

  

  ddale

  committed
  d96197a

  Copy full SHA for d96197a

• replace the warnings module with pytest; fix the plotting errors

  

  ddale

  committed
  08355a4

  Copy full SHA for 08355a4

Commits on Dec 13, 2020

• Merge branch 'master' into quantile-regression

  avidale

  authored
  fd37bec

  Copy full SHA for fd37bec

Commits on Mar 1, 2021

• Merge branch 'main' into quantile-regression

  avidale

  committed
  15e7198

  Copy full SHA for 15e7198

• update the code to make the old tests pass

  avidale

  committed
  9182213

  Copy full SHA for 9182213

• resolve most of the old comments

  avidale

  committed
  15e51ec

  Copy full SHA for 15e51ec

• update the way of checking training data

  avidale

  committed
  46fd074

  Copy full SHA for 46fd074

• remove unused imports

  avidale

  committed
  07b1faa

  Copy full SHA for 07b1faa

Commits on Mar 2, 2021

• replace the approximated nonlinear solver with the linear solver

  avidale

  committed
  0752966

  Copy full SHA for 0752966

Commits on Mar 14, 2021

• cosmetic changes to the interface of QuantileRegressor

  avidale

  committed
  97d3141

  Copy full SHA for 97d3141

Commits on Mar 15, 2021

• fix a couple of common tests, remove the normalization test

  avidale

  committed
  c31e038

  Copy full SHA for c31e038

• add a test for asymmetric error distribution; parametrize the tests

  avidale

  committed
  110e589

  Copy full SHA for 110e589

• update the example

  avidale

  committed
  7d83faa

  Copy full SHA for 7d83faa

• fix docstring parameters

  avidale

  committed
  b748dd5

  Copy full SHA for b748dd5

• add the feature to the changelog

  avidale

  committed
  6ea1057

  Copy full SHA for 6ea1057

• update the documentation for quantile regression

  avidale

  committed
  bba5018

  Copy full SHA for bba5018

• merge upstream

  avidale

  committed
  9b86fb4

  Copy full SHA for 9b86fb4

Commits on Mar 16, 2021

• use simplex method with older versions of scipy

  avidale

  committed
  3bcd2c2

  Copy full SHA for 3bcd2c2

Commits on Mar 17, 2021

• reuse sp_version

  avidale

  committed
  3e9b4b5

  Copy full SHA for 3e9b4b5

• skip the tests in case of old scipy

  avidale

  committed
  7006a79

  Copy full SHA for 7006a79

• Update doc/modules/linear_model.rst

  Show description for e67873e
  

  avidale

  and

  lorentzenchr

  authored
  e67873e

  Copy full SHA for e67873e

• Update doc/modules/linear_model.rst

  Show description for 23b8488
  

  avidale

  and

  lorentzenchr

  authored
  23b8488

  Copy full SHA for 23b8488

• Update doc/modules/linear_model.rst

  Show description for 4dafda5
  

  avidale

  and

  lorentzenchr

  authored
  4dafda5

  Copy full SHA for 4dafda5

• Update doc/modules/linear_model.rst

  Show description for 2284c56
  

  avidale

  and

  lorentzenchr

  authored
  2284c56

  Copy full SHA for 2284c56

• Update doc/modules/linear_model.rst

  Show description for 8e8c28d
  

  avidale

  and

  lorentzenchr

  authored
  8e8c28d

  Copy full SHA for 8e8c28d

• Update doc/modules/linear_model.rst

  Show description for a063325
  

  avidale

  and

  lorentzenchr

  authored
  a063325

  Copy full SHA for a063325

• move the diff to the 1.0 changelog

  avidale

  committed
  e546c8b

  Copy full SHA for e546c8b

• try even harder to avoid being tested with old scipy

  avidale

  committed
  08609a7

  Copy full SHA for 08609a7

• import ConvergenceWarning

  avidale

  committed
  3608130

  Copy full SHA for 3608130

• undo the exit from the example

  avidale

  committed
  3e16896

  Copy full SHA for 3e16896

Commits on Mar 22, 2021

• Update sklearn/linear_model/_quantile.py

  Show description for 6bb55ac
  

  avidale

  and

  lorentzenchr

  authored
  6bb55ac

  Copy full SHA for 6bb55ac

• Update sklearn/linear_model/_quantile.py

  Show description for 9a2b902
  

  avidale

  and

  lorentzenchr

  authored
  9a2b902

  Copy full SHA for 9a2b902

• Update sklearn/linear_model/_quantile.py

  Show description for 434419a
  

  avidale

  and

  lorentzenchr

  authored
  434419a

  Copy full SHA for 434419a

• Update sklearn/linear_model/_quantile.py

  Show description for 5ed4591
  

  avidale

  and

  lorentzenchr

  authored
  5ed4591

  Copy full SHA for 5ed4591

• Update sklearn/linear_model/_quantile.py

  Show description for 2a61dce
  

  avidale

  and

  lorentzenchr

  authored
  2a61dce

  Copy full SHA for 2a61dce

• Update sklearn/linear_model/_quantile.py

  Show description for 9059c42
  

  avidale

  and

  lorentzenchr

  authored
  9059c42

  Copy full SHA for 9059c42

• Update sklearn/linear_model/_quantile.py

  Show description for eacee73
  

  avidale

  and

  lorentzenchr

  authored
  eacee73

  Copy full SHA for eacee73

• Update sklearn/linear_model/_quantile.py

  Show description for 364c6b7
  

  avidale

  and

  lorentzenchr

  authored
  364c6b7

  Copy full SHA for 364c6b7

• Update sklearn/linear_model/_quantile.py

  Show description for 2107f1f
  

  avidale

  and

  lorentzenchr

  authored
  2107f1f

  Copy full SHA for 2107f1f

• Update sklearn/linear_model/_quantile.py

  Show description for ce8b268
  

  avidale

  and

  lorentzenchr

  authored
  ce8b268

  Copy full SHA for ce8b268

• Update sklearn/linear_model/_quantile.py

  Show description for a77ca8d
  

  avidale

  and

  lorentzenchr

  authored
  a77ca8d

  Copy full SHA for a77ca8d

• Update sklearn/linear_model/_quantile.py

  Show description for 03fceeb
  

  avidale

  and

  lorentzenchr

  authored
  03fceeb

  Copy full SHA for 03fceeb

• Update sklearn/linear_model/_quantile.py

  Show description for 6e8cdfd
  

  avidale

  and

  lorentzenchr

  authored
  6e8cdfd

  Copy full SHA for 6e8cdfd

• Update sklearn/linear_model/_quantile.py

  Show description for 4e49c9b
  

  avidale

  and

  lorentzenchr

  authored
  4e49c9b

  Copy full SHA for 4e49c9b

• Update sklearn/linear_model/_quantile.py

  Show description for 08999d0
  

  avidale

  and

  lorentzenchr

  authored
  08999d0

  Copy full SHA for 08999d0

• Update sklearn/linear_model/_quantile.py

  Show description for 52871fc
  

  avidale

  and

  lorentzenchr

  authored
  52871fc

  Copy full SHA for 52871fc

• Update sklearn/linear_model/_quantile.py

  Show description for 13a6327
  

  avidale

  and

  lorentzenchr

  authored
  13a6327

  Copy full SHA for 13a6327

• Update doc/modules/linear_model.rst

  Show description for 3902ce5
  

  avidale

  and

  lorentzenchr

  authored
  3902ce5

  Copy full SHA for 3902ce5

• Update doc/modules/linear_model.rst

  Show description for f9504b0
  

  avidale

  and

  lorentzenchr

  authored
  f9504b0

  Copy full SHA for f9504b0

Commits on Mar 29, 2021

• CLN use standard term variables n_samples and n_features

  lorentzenchr

  committed
  c16391d

  Copy full SHA for c16391d

• CLN rescale penalty

  lorentzenchr

  committed
  40f836e

  Copy full SHA for 40f836e

• CLN rename c_vector to c

  lorentzenchr

  committed
  f8ea953

  Copy full SHA for f8ea953

• DOC solver default argument

  lorentzenchr

  committed
  55d60d7

  Copy full SHA for 55d60d7

• CLN remove centering of y and X, i.e. _preprocess_data

  lorentzenchr

  committed
  6b67160

  Copy full SHA for 6b67160

• CLN restructure code

  Show description for 664e6f7
  lorentzenchr

  committed
  664e6f7

  Copy full SHA for 664e6f7

• TST parametrize test_quantile_estimates_fraction

  lorentzenchr

  committed
  7f73518

  Copy full SHA for 7f73518

• CLN add note about centering y and X

  lorentzenchr

  committed
  21715f2

  Copy full SHA for 21715f2

• DOC some more line breaks

  lorentzenchr

  committed
  4e23aae

  Copy full SHA for 4e23aae

• EXA fix title

  lorentzenchr

  committed
  f37e864

  Copy full SHA for f37e864

• TST rename test to test_quantile_estimates_calibration

  lorentzenchr

  committed
  7fb4a6b

  Copy full SHA for 7fb4a6b

• TST use pytest.raises instead of assert_raises

  lorentzenchr

  committed
  8f4e46b

  Copy full SHA for 8f4e46b

• TST fix test

  lorentzenchr

  committed
  452666f

  Copy full SHA for 452666f

• TST remove use of check_random_state

  lorentzenchr

  committed
  c37cf84

  Copy full SHA for c37cf84

• DOC update whatsnew

  lorentzenchr

  committed
  55cea88

  Copy full SHA for 55cea88

Commits on Mar 30, 2021

• TST nicer input validation tests

  lorentzenchr

  committed
  0262c32

  Copy full SHA for 0262c32

• CLN improve input validation of parameters

  Show description for f6128a7
  lorentzenchr

  committed
  f6128a7

  Copy full SHA for f6128a7

• TST improve test_asymmetric_error

  lorentzenchr

  committed
  4d33624

  Copy full SHA for 4d33624

• TST add test with L1 and compare to Nelder-Mead

  lorentzenchr

  committed
  857f281

  Copy full SHA for 857f281

• TST parametrize test_quantile_equals_huber_for_low_epsilon

  lorentzenchr

  committed
  be06252

  Copy full SHA for be06252

• TST use pytest.approx

  lorentzenchr

  committed
  45fe6a3

  Copy full SHA for 45fe6a3

Commits on Mar 31, 2021

• CLN fix type

  lorentzenchr

  committed
  5653caa

  Copy full SHA for 5653caa

• DOC remove 'auto' solver

  lorentzenchr

  committed
  b046ba3

  Copy full SHA for b046ba3

• TST change order of pytest.mark decorators

  lorentzenchr

  committed
  54e21a8

  Copy full SHA for 54e21a8

Commits on Apr 1, 2021

• DOC change references to quantile regression literature

  lorentzenchr

  committed
  33e766c

  Copy full SHA for 33e766c

• TST add test for equivariance

  lorentzenchr

  committed
  9541e91

  Copy full SHA for 9541e91

• DOC miiiinor improvements

  lorentzenchr

  committed
  cb65aae

  Copy full SHA for cb65aae

• DOC recommend highs methods

  lorentzenchr

  committed
  29995cf

  Copy full SHA for 29995cf

Commits on Apr 2, 2021

• CLN make init arguments keyword only

  lorentzenchr

  committed
  03c6255

  Copy full SHA for 03c6255

Commits on Apr 18, 2021

• ENH remove simplex solver

  Show description for 1c6aa97
  lorentzenchr

  committed
  1c6aa97

  Copy full SHA for 1c6aa97

• Merge branch 'main' into quantile-regression

  lorentzenchr

  committed
  b07bb9a

  Copy full SHA for b07bb9a

Commits on Apr 26, 2021

• Add and fix author names

  lorentzenchr

  committed
  3d22545

  Copy full SHA for 3d22545

• Change default alpha to 1.0

  lorentzenchr

  committed
  0fb75d1

  Copy full SHA for 0fb75d1

• Use f-strings and string fixes

  lorentzenchr

  committed
  0265d0e

  Copy full SHA for 0265d0e

• Better docstring for solver highs

  lorentzenchr

  committed
  1667eaf

  Copy full SHA for 1667eaf

• Prefer np.full over np.ones

  lorentzenchr

  committed
  50f6eaf

  Copy full SHA for 50f6eaf

• Better warning message

  lorentzenchr

  committed
  4a966fe

  Copy full SHA for 4a966fe

• Change warning into error on failing linprog

  lorentzenchr

  committed
  c076490

  Copy full SHA for c076490

• TST add test for failing linprog

  lorentzenchr

  committed
  e06ee29

  Copy full SHA for e06ee29

• Merge branch 'main' into quantile-regression

  lorentzenchr

  committed
  56ff1c6

  Copy full SHA for 56ff1c6

Commits on Apr 27, 2021

• DOC add See Also section

  lorentzenchr

  committed
  4e62a19

  Copy full SHA for 4e62a19

• Address review comments

  lorentzenchr

  committed
  1c1f926

  Copy full SHA for 1c1f926

Commits on Apr 30, 2021

• EXA rework example

  lorentzenchr

  committed
  dc63a8b

  Copy full SHA for dc63a8b

• make linter happy

  lorentzenchr

  committed
  1109456

  Copy full SHA for 1109456

• EXA tiny improvement

  lorentzenchr

  committed
  23ca0fc

  Copy full SHA for 23ca0fc

Commits on May 6, 2021

• TST try to fix two tests

  lorentzenchr

  committed
  84da276

  Copy full SHA for 84da276

• FIX zero samples weights

  lorentzenchr

  committed
  23b5a22

  Copy full SHA for 23b5a22

• make linter happy

  lorentzenchr

  committed
  eff8b12

  Copy full SHA for eff8b12

Commits on May 7, 2021

• TST add test for solver_options

  lorentzenchr

  committed
  0fc55e2

  Copy full SHA for 0fc55e2

• ENH make solver more stable (but slower)

  lorentzenchr

  committed
  1aed18e

  Copy full SHA for 1aed18e

• black

  lorentzenchr

  committed
  354548b

  Copy full SHA for 354548b

• fix solver_option logic

  lorentzenchr

  committed
  ec672f5

  Copy full SHA for ec672f5

• make linter happy again

  lorentzenchr

  committed
  c3a520f

  Copy full SHA for c3a520f

Commits on May 18, 2021

• Merge branch 'main' into quantile-regression

  lorentzenchr

  committed
  0278f2d

  Copy full SHA for 0278f2d

• TST adapt test precision

  lorentzenchr

  committed
  2057e03

  Copy full SHA for 2057e03

• ENH print linprog result.message when failing

  lorentzenchr

  committed
  0c5f220

  Copy full SHA for 0c5f220

• TST assure convergence in test_equivariance

  lorentzenchr

  committed
  1f244e1

  Copy full SHA for 1f244e1

• TST remove skipif for scipy<1.0.0

  lorentzenchr

  committed
  0701b62

  Copy full SHA for 0701b62

• MNT remove _more_tags

  lorentzenchr

  committed
  45e4b5a

  Copy full SHA for 45e4b5a

Commits on May 20, 2021

• prefer ConvergenceWarning over ValueError

  lorentzenchr

  committed
  3cca6db

  Copy full SHA for 3cca6db

• DOC add examples section in docstring

  lorentzenchr

  committed
  346eba5

  Copy full SHA for 346eba5

• address minor review comments

  lorentzenchr

  committed
  5b7446d

  Copy full SHA for 5b7446d

• TST test highs solver raises for old scipy

  lorentzenchr

  committed
  14d00d4

  Copy full SHA for 14d00d4

• fix use of fixture

  lorentzenchr

  committed
  38c0845

  Copy full SHA for 38c0845

• fix example by calling fit

  lorentzenchr

  committed
  0d064cf

  Copy full SHA for 0d064cf

Commits on May 23, 2021

• DOC add reference to Koenker's book

  lorentzenchr

  committed
  6f870a4

  Copy full SHA for 6f870a4