Appendix D — Errata

This appendix lists changes to the online version of this book to chapters included in the first edition.

D.1 1. Introduction and Overview

Added note about Docker images.

D.2 2. Data and Basic Modeling

Replaced reference to Param with Domain.
Added paragraph about $configure() method.
Replaced the PimaIndiansDiabetes2 data set in the exercises with the SynthDiabetes2 data set, because mlbench removed PimaIndiansDiabetes2.

D.3 3. Evaluation and Benchmarking

Use $encapsulate() method instead of the $encapsulate and $fallback fields.
A section on the mlr3inferr package was added.

D.4 4. Hyperparameter Optimization

Renamed TuningInstanceSingleCrit to TuningInstanceBatchSingleCrit.
Renamed TuningInstanceMultiCrit to TuningInstanceBatchMultiCrit.
Renamed Tuner to TunerBatch.
Replaced reference to Param with Domain.
Replace lrn("surv.coxtime") with lrn("classif.mlp").
Added note that learner dependencies are automatically preserved when using to_tune().
Added booster = "gbtree" to the lrn("classif.xgboost") example.

D.5 5. Advanced Tuning Methods and Black Box Optimization

Renamed TuningInstanceSingleCrit to TuningInstanceBatchSingleCrit.
Renamed TuningInstanceMultiCrit to TuningInstanceBatchMultiCrit.
Renamed Tuner to TunerBatch.
Renamed OptimInstanceSingleCrit to OptimInstanceBatchSingleCrit.
Renamed OptimInstanceMultiCrit to OptimInstanceBatchMultiCrit.
Renamed Optimizer to OptimizerBatch.
Replaced OptimInstanceSingleCrit$new() with oi().
Add oi() to the table about important functions.
Use $encapsulate() method instead of the $encapsulate and $fallback fields.
In example 5.4.4 lrn("svm") was tuned instead of lrn("rpart").

D.6 6. Feature Selection

Renamed FSelectInstanceSingleCrit to FSelectInstanceBatchSingleCrit.
Renamed FSelectInstanceMultiCrit to FSelectInstanceBatchMultiCrit.
Renamed FeatureSelector to FeatureSelectorBatch.
Add fsi() to the table about important functions.

D.7 7. Sequential Pipelines

Replaced tsk("pima") with tsk("diabetes"), because the pima task was removed from mlr3.

D.8 8. Non-sequential Pipelines and Tuning

Reduce the number of cores to 2 in the chunking example.
Use $encapsulate() method instead of the $encapsulate and $fallback fields.
Replaced tsk("pima") with tsk("diabetes"), because the pima task was removed from mlr3.
Replaced pos() with named lists of PipeOps in the AutoML exercise solution, because pos() now returns an unnamed list and ppl("branch") would otherwise use integer branch selection values.

D.9 10. Advanced Technical Aspects of mlr3

Use $encapsulate() method instead of the $encapsulate and $fallback fields.
Add section on parallelization with mirai package.
Added section on condition classes.
Added section on base logger.

D.10 11. Large-Scale Benchmarking

Use $encapsulate() method instead of the $encapsulate and $fallback fields.
Replaced tsk("pima") with tsk("diabetes"), because the pima task was removed from mlr3.

D.11 12. Model Interpretation

Subset task to row 127 instead of 35 for the local surrogate model.
Add as.data.frame() to “Correctly Interpreting Shapley Values” section.

D.12 13. Beyond Regression and Classification

Use gamma instead of gamma.mu for lrn("surv.svm")
Substitute RCLL with ISBS measure
Mention pipeline_responsecompositor() pipeline for changing predict types
Use lrn("surv.xgboost.aft") instead of lrn("surv.glmnet") in “Composition” subsection
Added a note about pendensity being removed from CRAN in 2026.

D.13 14. Algorithmic Fairness

Added a seed to code chunk 7.

--- aliases: - "/errata.html" --- # Errata {#sec-appendix-errata} ::: {.content-visible when-format="html"} {{< include ../../common/_setup.qmd >}} This appendix lists changes to the online version of this book to chapters included in the first edition. ## 1. Introduction and Overview * Added note about Docker images. ## 2. Data and Basic Modeling * Replaced reference to `Param` with `Domain`. * Added paragraph about `$configure()` method. * Replaced the `PimaIndiansDiabetes2` data set in the exercises with the `SynthDiabetes2` data set, because `mlbench` removed `PimaIndiansDiabetes2`. ## 3. Evaluation and Benchmarking * Use `$encapsulate()` method instead of the `$encapsulate` and `$fallback` fields. * A section on the `mlr3inferr` package was added. ## 4. Hyperparameter Optimization * Renamed `TuningInstanceSingleCrit` to `TuningInstanceBatchSingleCrit`. * Renamed `TuningInstanceMultiCrit` to `TuningInstanceBatchMultiCrit`. * Renamed `Tuner` to `TunerBatch`. * Replaced reference to `Param` with `Domain`. * Replace `lrn("surv.coxtime")` with `lrn("classif.mlp")`. * Added note that learner dependencies are automatically preserved when using `to_tune()`. * Added `booster = "gbtree"` to the `lrn("classif.xgboost")` example. ## 5. Advanced Tuning Methods and Black Box Optimization * Renamed `TuningInstanceSingleCrit` to `TuningInstanceBatchSingleCrit`. * Renamed `TuningInstanceMultiCrit` to `TuningInstanceBatchMultiCrit`. * Renamed `Tuner` to `TunerBatch`. * Renamed `OptimInstanceSingleCrit` to `OptimInstanceBatchSingleCrit`. * Renamed `OptimInstanceMultiCrit` to `OptimInstanceBatchMultiCrit`. * Renamed `Optimizer` to `OptimizerBatch`. * Replaced `OptimInstanceSingleCrit$new()` with `oi()`. * Add `oi()` to the table about important functions. * Use `$encapsulate()` method instead of the `$encapsulate` and `$fallback` fields. * In example 5.4.4 `lrn("svm")` was tuned instead of `lrn("rpart")`. ## 6. Feature Selection * Renamed `FSelectInstanceSingleCrit` to `FSelectInstanceBatchSingleCrit`. * Renamed `FSelectInstanceMultiCrit` to `FSelectInstanceBatchMultiCrit`. * Renamed `FeatureSelector` to `FeatureSelectorBatch`. * Add `fsi()` to the table about important functions. ## 7. Sequential Pipelines * Replaced `tsk("pima")` with `tsk("diabetes")`, because the `pima` task was removed from `mlr3`. ## 8. Non-sequential Pipelines and Tuning * Reduce the number of cores to 2 in the chunking example. * Use `$encapsulate()` method instead of the `$encapsulate` and `$fallback` fields. * Replaced `tsk("pima")` with `tsk("diabetes")`, because the `pima` task was removed from `mlr3`. * Replaced `pos()` with named lists of `PipeOp`s in the AutoML exercise solution, because `pos()` now returns an unnamed list and `ppl("branch")` would otherwise use integer branch selection values. ## 10. Advanced Technical Aspects of mlr3 * Use `$encapsulate()` method instead of the `$encapsulate` and `$fallback` fields. * Add section on parallelization with `mirai` package. * Added section on condition classes. * Added section on base logger. ## 11. Large-Scale Benchmarking * Use `$encapsulate()` method instead of the `$encapsulate` and `$fallback` fields. * Replaced `tsk("pima")` with `tsk("diabetes")`, because the `pima` task was removed from `mlr3`. ## 12. Model Interpretation * Subset task to row 127 instead of 35 for the local surrogate model. * Add `as.data.frame()` to "Correctly Interpreting Shapley Values" section. ## 13. Beyond Regression and Classification * Use `gamma` instead of `gamma.mu` for `lrn("surv.svm")` * Substitute RCLL with ISBS measure * Mention `pipeline_responsecompositor()` pipeline for changing predict types * Use `lrn("surv.xgboost.aft")` instead of `lrn("surv.glmnet")` in "Composition" subsection * Added a note about `pendensity` being removed from CRAN in 2026. ## 14. Algorithmic Fairness * Added a seed to code chunk 7.