You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* add a quantile regression mode to test with
* update type checkers
* avoid confusion with global all_models object
* add quantile_level argument to set_mode()
* initial data for quantreg
* some initial tests
* fix some issues
* enable quantile prediction
* tests for quantreg
* Quantile predictions output constructor (#1191)
* small change to predict checks
* add vctrs for quantiles and test, refactor *_rq_preds
* revise tests
* Apply some of the suggestions from code review
Co-authored-by: Simon P. Couch <simonpatrickcouch@gmail.com>
* rename tests on suggestion from code review
* export missing funs from vctrs for formatting
* convert errors to snapshot tests
* pass call through input check
* update snapshots for caller_env
* rename to parsnip_quantiles, add format snapshot tests
* Apply suggestions from @topepo
Co-authored-by: Max Kuhn <mxkuhn@gmail.com>
* rename parsnip_quantiles to quantile_pred
* rename parsnip_quantiles to quantile_pred and add vector probability check
* fix: two bugs introduced earlier
* add formatting tests for single quantile
* replace walk with a loop to avoid "Error in map()"
* remove row/col names
* adjust quantile_pred format
* as_tibble method
* updated NEWS file
* add PR number
* small new update
* helper methods
* update docs
* re-enable quantiles prediction for #1203
* update some tests
* no longer needed
* use tibble::new_tibble
* braces
* test as_tibble
* remove print methods
---------
Co-authored-by: Simon P. Couch <simonpatrickcouch@gmail.com>
Co-authored-by: Max Kuhn <mxkuhn@gmail.com>
Co-authored-by: ‘topepo’ <‘mxkuhn@gmail.com’>
* quantile regression updates for new hardhat model (#1207)
* bump hardhat version
* remove parts now in hardhat
* update for new hardhat version
* quantile_levels (plural now)
* news update
* typo
* rename helper function
* run CI on PRs from branches
* forgotten remote
* actions for edited PRs
* plural
* expand branch list
* export function for censored to use
* updated snapshot
* remake snapshot
* Revert "remake snapshot"
This reverts commit 954e326.
* updated snapshot
* Update R/arguments.R
Co-authored-by: Hannah Frick <hfrick@users.noreply.github.com>
* typo
* changes from reviewer feedback
---------
Co-authored-by: ‘topepo’ <‘mxkuhn@gmail.com’>
Co-authored-by: Hannah Frick <hfrick@users.noreply.github.com>
* Change to `quantile` argument to `quantile levels` (#1208)
* quantile -> quantile_levels for #1203
* defer test until censored updates in new PR
* update docs for quantile_levels
* update test
* disable quantile predictions for surv_reg
---------
Co-authored-by: ‘topepo’ <‘mxkuhn@gmail.com’>
* post conflict merge updates
* update news
* version bump and fix typo
* revert GHA branches
* small bug fix
* Apply suggestions from code review
Co-authored-by: Hannah Frick <hfrick@users.noreply.github.com>
Co-authored-by: Emil Hvitfeldt <emil.hvitfeldt@posit.co>
* don't export median
* add call arg
* added documentation on model
* add mode
* convert error to warning
* remove rankdeficient
* added skip
* add deprecated `quantile` arg back in
* remove numeric prediction
---------
Co-authored-by: ‘topepo’ <‘mxkuhn@gmail.com’>
Co-authored-by: Daniel McDonald <dajmcdon@gmail.com>
Co-authored-by: Simon P. Couch <simonpatrickcouch@gmail.com>
Co-authored-by: Hannah Frick <hfrick@users.noreply.github.com>
Co-authored-by: Emil Hvitfeldt <emil.hvitfeldt@posit.co>
Copy file name to clipboardExpand all lines: NEWS.md
+20-4
Original file line number
Diff line number
Diff line change
@@ -1,25 +1,41 @@
1
1
# parsnip (development version)
2
2
3
+
## New Features
4
+
5
+
* A new model mode (`"quantile regression"`) was added. Including:
6
+
* A `linear_reg()` engine for `"quantreg"`.
7
+
* Predictions are encoded via a custom vector type. See [hardhat::quantile_pred()].
8
+
* Predicted quantile levels are designated when the new mode is specified. See `?set_mode`.
9
+
3
10
*`fit_xy()` can now take dgCMatrix input for `x` argument (#1121).
4
11
5
12
*`fit_xy()` can now take sparse tibbles as data values (#1165).
6
13
7
14
*`predict()` can now take dgCMatrix and sparse tibble input for `new_data` argument, and error informatively when model doesn't support it (#1167).
8
15
9
-
* Transitioned package errors and warnings to use cli (#1147 and #1148 by
10
-
@shum461, #1153 by @RobLBaker and @wright13, #1154 by @JamesHWade, #1160,
11
-
#1161, #1081).
16
+
* New `extract_fit_time()` method has been added that returns the time it took to train the model (#853).
17
+
18
+
## Other Changes
19
+
20
+
* Transitioned package errors and warnings to use cli (#1147 and #1148 by @shum461, #1153 by @RobLBaker and @wright13, #1154 by @JamesHWade, #1160, #1161, #1081).
12
21
13
22
*`fit_xy()` currently raises an error for `gen_additive_mod()` model specifications as the default engine (`"mgcv"`) specifies smoothing terms in model formulas. However, some engines specify smooths via additional arguments, in which case the restriction on `fit_xy()` is excessive. parsnip will now only raise an error when fitting a `gen_additive_mod()` with `fit_xy()` when using the `"mgcv"` engine (#775).
14
23
15
24
* Aligned `null_model()` with other model types; the model type now has an engine argument that defaults to `"parsnip"` and is checked with the same machinery that checks other model types in the package (#1083).
16
25
17
-
* New `extract_fit_time()` method has been added that returns the time it took to train the model (#853).
26
+
## Bug Fixes
18
27
19
28
* Ensure that `knit_engine_docs()` has the required packages installed (#1156).
20
29
21
30
* Fixed bug where some models fit using `fit_xy()` couldn't predict (#1166).
22
31
32
+
## Breaking Change
33
+
34
+
* For quantile prediction, the `quantile` argument to `predict()` has been deprecate in facor of `quantile_levels`. This does not affect models with mode `"quantile regression"`.
35
+
36
+
* The quantile regression prediction type was disabled for the deprecated `surv_reg()` model.
37
+
38
+
23
39
# parsnip 1.2.1
24
40
25
41
* Added a missing `tidy()` method for survival analysis glmnet models (#1086).
0 commit comments