Merge branch 'master' into s3params_compat

.github/dependabot.yml

@@ -0,0 +1,6 @@
+version: 2
+updates:
+  - package-ecosystem: "github-actions"
+    directory: "/"
+    schedule:
+      interval: "weekly"

.github/workflows/dev-cmd-check.yml

@@ -27,7 +27,7 @@ jobs:
           - {os: ubuntu-latest, r: 'release', dev-package: 'mlr-org/mlr3'}
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-pandoc@v2
 

.github/workflows/pkgdown.yml

@@ -23,7 +23,7 @@ jobs:
     env:
       GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-pandoc@v2
 
@@ -44,7 +44,7 @@ jobs:
 
       - name: Deploy
         if: github.event_name != 'pull_request'
-        uses: JamesIves/github-pages-deploy-action@v4.4.1
+        uses: JamesIves/github-pages-deploy-action@v4.5.0
         with:
           clean: false
           branch: gh-pages

.github/workflows/r-cmd-check.yml

@@ -28,7 +28,7 @@ jobs:
           - {os: ubuntu-latest,   r: 'release'}
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-pandoc@v2
 

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: mlr3proba
 Title: Probabilistic Supervised Learning for 'mlr3'
-Version: 0.5.8
+Version: 0.6.0
 Authors@R:
     c(person(given = "Raphael",
              family = "Sonabend",
@@ -79,12 +79,12 @@ Encoding: UTF-8
 LazyData: true
 NeedsCompilation: no
 Roxygen: list(markdown = TRUE, r6 = TRUE)
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.1.9000
 Collate:
     'LearnerDens.R'
+    'aaa.R'
     'LearnerDensHistogram.R'
     'LearnerDensKDE.R'
-    'LearnerRegrGaussian.R'
     'LearnerSurv.R'
     'LearnerSurvCoxPH.R'
     'LearnerSurvKaplan.R'
@@ -135,11 +135,11 @@ Collate:
     'PredictionSurv.R'
     'RcppExports.R'
     'TaskDens.R'
-    'TaskDens_X.R'
+    'TaskDens_zzz.R'
     'TaskGeneratorSimdens.R'
     'TaskGeneratorSimsurv.R'
     'TaskSurv.R'
-    'TaskSurv_X.R'
+    'TaskSurv_zzz.R'
     'as_prediction_dens.R'
     'as_prediction_surv.R'
     'as_task_dens.R'

NAMESPACE

@@ -93,8 +93,6 @@ export(as_task_dens)
 export(as_task_surv)
 export(assert_surv)
 export(breslow)
-export(crankcompositor)
-export(distrcompositor)
 export(pecs)
 export(pipeline_survtoregr)
 export(plot_probregr)
@@ -118,6 +116,7 @@ importFrom(stats,reformulate)
 importFrom(stats,sd)
 importFrom(survival,Surv)
 importFrom(utils,data)
+importFrom(utils,getFromNamespace)
 importFrom(utils,head)
 importFrom(utils,tail)
 useDynLib(mlr3proba)

NEWS.md

@@ -1,3 +1,14 @@
+# mlr3proba 0.6.0
+
+* Optimized `surv.logloss` and `calib_alpha` measures (bypassing `distr6`)
+* Update/refine all measure docs (namign conventions from upcoming scoring rules paper) + doc templates
+* fix very rare bugs in `calib_alpha`, `surv.logloss` and `surv.graf` (version with proper = FALSE)
+
+# mlr3proba 0.5.9
+
+* Fix several old issues (#348, #301, #281)
+* `distrcompositor` and `crankcompositor` deprecated functions were completely removed
+
 # mlr3proba 0.5.8
 
 * Fix Breslow issues (`phash` warning and `Inf` lp predictions)

R/LearnerDensHistogram.R

@@ -38,3 +38,6 @@ LearnerDensHistogram = R6::R6Class("LearnerDensHistogram",
     }
   )
 )
+
+#' @include aaa.R
+register_learner("dens.hist", LearnerDensHistogram)

R/LearnerDensKDE.R

@@ -1,7 +1,7 @@
 #' @template dens_learner
 #' @templateVar title Kernel
 #' @templateVar fullname LearnerDensKDE
-#' @templateVar caller kernels implemented in \CRANpkg{distr6}
+#' @templateVar caller kernels implemented in [distr6]
 #' @details The default bandwidth uses Silverman's rule-of-thumb for Gaussian kernels, however for
 #' non-Gaussian kernels it is recommended to use \CRANpkg{mlr3tuning} to tune the bandwidth with
 #' cross-validation. Other density learners can be used for automated bandwidth selection.
@@ -88,3 +88,5 @@ LearnerDensKDE = R6::R6Class("LearnerDensKDE",
     }
   )
 )
+
+register_learner("dens.kde", LearnerDensKDE)

R/LearnerSurv.R

@@ -7,7 +7,7 @@
 #' * Creates [Prediction]s of class [PredictionSurv].
 #' * Possible values for `predict_types` are:
 #'   - `"distr"`: Predicts a probability distribution for each observation in the test set,
-#'                uses \CRANpkg{distr6}.
+#'                uses [distr6].
 #'   - `"lp"`: Predicts a linear predictor for each observation in the test set.
 #'   - `"crank"`: Predicts a continuous ranking for each observation in the test set.
 #'   - `"response"`: Predicts a survival time for each observation in the test set.

R/LearnerSurvCoxPH.R

@@ -1,11 +1,10 @@
-#' @template surv_learner
 #' @templateVar title Cox Proportional Hazards
 #' @templateVar fullname LearnerSurvCoxPH
 #' @templateVar caller [survival::coxph()]
 #' @templateVar distr by [survival::survfit.coxph()]
 #' @templateVar lp by [survival::predict.coxph()]
-#' @description
-#'
+#' @templateVar id surv.coxph
+#' @template surv_learner
 #'
 #' @references
 #' `r format_bib("cox_1972")`
@@ -71,3 +70,5 @@ LearnerSurvCoxPH = R6Class("LearnerSurvCoxPH",
     }
   )
 )
+
+register_learner("surv.coxph", LearnerSurvCoxPH)

R/LearnerSurvKaplan.R

@@ -1,11 +1,9 @@
-#' @template surv_learner
 #' @templateVar title Kaplan-Meier Estimator
 #' @templateVar fullname LearnerSurvKaplan
 #' @templateVar caller [survival::survfit()]
 #' @templateVar distr by estimating the survival function with [survival::survfit()]
-#'
-#' @description
-#'
+#' @templateVar id surv.kaplan
+#' @template surv_learner
 #'
 #' @references
 #' `r format_bib("kaplan_1958")`
@@ -35,13 +33,13 @@ LearnerSurvKaplan = R6Class("LearnerSurvKaplan",
     },
 
     .predict = function(task) {
-
-      # Define WeightedDiscrete distr6 distribution from the survival function
       times = self$model$time
-      surv = matrix(rep(self$model$surv, task$nrow), ncol = length(times), nrow = task$nrow,
-        byrow = TRUE)
+      surv = matrix(rep(self$model$surv, task$nrow), ncol = length(times),
+                    nrow = task$nrow, byrow = TRUE)
 
       .surv_return(times = times, surv = surv)
     }
   )
 )
+
+register_learner("surv.kaplan", LearnerSurvKaplan)

R/LearnerSurvRpart.R

@@ -1,12 +1,14 @@
-#' @template surv_learner
 #' @templateVar title Rpart Survival Trees
 #' @templateVar fullname LearnerSurvRpart
 #' @templateVar caller [rpart::rpart()]
 #' @templateVar crank using [rpart::predict.rpart()]
+#' @templateVar id surv.rpart
+#' @template surv_learner
 #'
-#' @description
-#' Parameter `xval` is set to 0 in order to save some computation time.
-#' Parameter `model` has been renamed to `keep_model`.
+#' @section Initial parameter values:
+#'
+#' - `xval` is set to 0 in order to save some computation time.
+#' - `model` has been renamed to `keep_model`.
 #'
 #' @references
 #' `r format_bib("breiman_1984")`
@@ -87,3 +89,5 @@ LearnerSurvRpart = R6Class("LearnerSurvRpart",
     }
   )
 )
+
+register_learner("surv.rpart", LearnerSurvRpart)

R/MeasureDensLogloss.R

@@ -1,16 +1,15 @@
 #' @template dens_measure
-#' @templateVar title Log loss
+#' @templateVar title Log Loss
 #' @templateVar inherit [MeasureDens]
 #' @templateVar fullname MeasureDensLogloss
-#' @templateVar pars eps = 1e-15
-#' @templateVar eps_par TRUE
-#'
+#' @templateVar eps 1e-15
 #' @template param_eps
 #'
 #' @description
 #' Calculates the cross-entropy, or logarithmic (log), loss.
 #'
-#' The logloss, in the context of probabilistic predictions, is defined as the negative log
+#' @details
+#' The Log Loss, in the context of probabilistic predictions, is defined as the negative log
 #' probability density function, \eqn{f}, evaluated at the observed value, \eqn{y},
 #' \deqn{L(f, y) = -\log(f(y))}{L(f, y) = -log(f(y))}
 #'
@@ -47,3 +46,5 @@ MeasureDensLogloss = R6::R6Class("MeasureDensLogloss",
     }
   )
 )
+
+register_measure("dens.logloss", MeasureDensLogloss)

R/MeasureRegrLogloss.R

@@ -1,16 +1,15 @@
 #' @template regr_measure
-#' @templateVar title Log loss
+#' @templateVar title Log Loss
 #' @templateVar inherit [MeasureRegr]
 #' @templateVar fullname MeasureRegrLogloss
-#' @templateVar pars eps = 1e-15
-#' @templateVar eps_par TRUE
-#'
+#' @templateVar eps 1e-15
 #' @template param_eps
 #'
 #' @description
 #' Calculates the cross-entropy, or logarithmic (log), loss.
 #'
-#' The logloss, in the context of probabilistic predictions, is defined as the negative log
+#' @details
+#' The Log Loss, in the context of probabilistic predictions, is defined as the negative log
 #' probability density function, \eqn{f}, evaluated at the observed value, \eqn{y},
 #' \deqn{L(f, y) = -\log(f(y))}{L(f, y) = -log(f(y))}
 #'
@@ -54,3 +53,5 @@ MeasureRegrLogloss = R6::R6Class("MeasureRegrLogloss",
     }
   )
 )
+
+register_measure("regr.logloss", MeasureRegrLogloss)

R/MeasureSurv.R

@@ -21,7 +21,6 @@
 #' @template param_packages
 #' @template param_label
 #' @template param_man
-#' @template param_se
 #'
 #' @family Measure
 #' @seealso
@@ -32,6 +31,8 @@ MeasureSurv = R6Class("MeasureSurv",
   public = list(
     #' @description
     #' Creates a new instance of this [R6][R6::R6Class] class.
+    #' @param se If `TRUE` then returns standard error of the
+    #' measure otherwise returns the mean (default).
     initialize = function(id, param_set = ps(), range, minimize = NA, aggregator = NULL,
       properties = character(), predict_type = "distr", task_properties = character(),
       packages = character(), label = NA_character_, man = NA_character_, se = FALSE) {