For a 0.20.7 release

Project.toml

@@ -1,7 +1,7 @@
 name = "MLJ"
 uuid = "add582a8-e3ab-11e8-2d5e-e98b27df1bc7"
 authors = ["Anthony D. Blaom <[email protected]>"]
-version = "0.20.6"
+version = "0.20.7"
 
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
@@ -32,9 +32,9 @@ Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 CategoricalArrays = "0.8,0.9, 0.10"
 ComputationalResources = "0.3"
 Distributions = "0.21,0.22,0.23, 0.24, 0.25"
-FeatureSelection = "0.1.1"
+FeatureSelection = "0.2"
 MLJBalancing = "0.1"
-MLJBase = "1"
+MLJBase = "1.5"
 MLJEnsembles = "0.4"
 MLJFlow = "0.5"
 MLJIteration = "0.6"

docs/ModelDescriptors.toml

@@ -154,6 +154,7 @@ MultitargetNeuralNetworkRegressor_BetaML = ["regression", "neural_networks"]
 MultitargetNeuralNetworkRegressor_MLJFlux = ["regression", "iterative_models", "neural_networks"]
 MultitargetRidgeRegressor_MultivariateStats = ["regression"]
 MultitargetSRRegressor_SymbolicRegression = ["regression"]
+NeuralNetworkBinaryClassifier_MLJFlux = ["classification", "neural_networks"]
 NeuralNetworkClassifier_BetaML = ["classification", "neural_networks"]
 NeuralNetworkClassifier_MLJFlux = ["classification", "iterative_models", "neural_networks"]
 NeuralNetworkRegressor_BetaML = ["regression", "neural_networks"]

docs/make.jl

@@ -67,10 +67,11 @@ pages = [
         "Generating Synthetic Data" => "generating_synthetic_data.md",
         "OpenML Integration" => "openml_integration.md",
     ],
-    "Model Basics" => [
+    "Models" => [
         "Model Search" => "model_search.md",
         "Loading Model Code" => "loading_model_code.md",
         "Transformers and Other Unsupervised models" => "transformers.md",
+        "Feature Selection" => "feature_selection.md",
         "List of Supported Models" => "list_of_supported_models.md",
     ],
     "Meta-algorithms" => [

docs/src/feature_selection.md

@@ -0,0 +1,13 @@
+# Feature Selection
+
+For more on feature selection tools, refer to the [FeatureSelection.jl](https://juliaai.github.io/FeatureSelection.jl/dev/) documentation.
+
+- [`FeatureSelector`](@ref)
+- [`RecursiveFeatureElimination`](@ref)
+
+## Reference
+
+```@docs
+FeatureSelector
+RecursiveFeatureElimination
+```

docs/src/model_search.md

@@ -1,11 +1,10 @@
 # [Model Search](@id model_search)
 
-MLJ has a model registry, allowing the user to search models and their
-properties, without loading all the packages containing model code. In
-turn, this allows one to efficiently find all models solving a given
-machine learning task. The task itself is specified with the help of
-the `matching` method, and the search executed with the `models`
-methods, as detailed below.
+In addition to perusing the [Model Browser](@ref), one can programatically search MLJ's
+Model Registry, without actually loading all the packages providing model code. This
+allows you to efficiently find all models solving a given machine learning task. The task
+itself is specified with the help of the `matching` method, and the search executed with
+the `models` methods, as detailed below.
 
 For commonly encountered problems with model search, see also
 [Preparing Data](@ref).

docs/src/transformers.md

@@ -24,18 +24,16 @@ that learn a probability distribution](@ref) for an example.
 
 ## Built-in transformers
 
-```@docs
-MLJModels.Standardizer
-MLJModels.OneHotEncoder
-MLJModels.ContinuousEncoder
-MLJModels.FillImputer
-MLJModels.UnivariateFillImputer
-FeatureSelection.FeatureSelector
-MLJModels.UnivariateBoxCoxTransformer
-MLJModels.UnivariateDiscretizer
-MLJModels.UnivariateTimeTypeToContinuous
-```
-
+- [`Standardizer`](@ref)
+- [`OneHotEncoder`](@ref)
+- [`ContinuousEncoder`](@ref)
+- [`FillImputer`](@ref)
+- [`UnivariateFillImputer`](@ref)
+- [`UnivariateBoxCoxTransformer`](@ref)
+- [`InteractionTransformer`](@ref)
+- [`UnivariateDiscretizer`](@ref)
+- [`UnivariateTimeTypeToContinuous`](@ref)
+- [`FeatureSelector`](@ref).
 
 ## Static transformers
 
@@ -231,3 +229,17 @@ compare[51:58]
 ```@example predtrans
 compare[101:108]
 ```
+
+## Reference
+
+```@docs
+MLJModels.Standardizer
+MLJModels.OneHotEncoder
+MLJModels.ContinuousEncoder
+MLJModels.FillImputer
+MLJModels.UnivariateFillImputer
+MLJModels.UnivariateBoxCoxTransformer
+MLJModels.InteractionTransformer
+MLJModels.UnivariateDiscretizer
+MLJModels.UnivariateTimeTypeToContinuous
+```

docs/src/working_with_categorical_data.md

@@ -345,3 +345,9 @@ d_vec = UnivariateFinite(["no", "yes"], yes_probs, augment=true, pool=v)
 ```
 
 For more options, see [`UnivariateFinite`](@ref).
+
+## Reference
+
+```@docs
+UnivariateFinite
+```

src/MLJ.jl

@@ -146,7 +146,7 @@ export nrows, color_off, color_on,
     @pipeline, Stack, Pipeline, TransformedTargetModel,
     ResamplingStrategy, Holdout, CV, TimeSeriesCV,
     StratifiedCV, evaluate!, Resampler, iterator, PerformanceEvaluation,
-    default_resource, pretty,
+    default_resource, default_logger, pretty,
     make_blobs, make_moons, make_circles, make_regression,
     fit_only!, return!, int, decoder,
     default_scitype_check_level,

test/integration.jl

@@ -155,7 +155,11 @@ PATHOLOGIES = filter(MODELS) do model
         # too slow to train!
         (model.name == "LOCIDetector" && model.package_name == "OutlierDetectionPython") ||
         # TO REDUCE TESTING TIME
-        model.package_name == "MLJScikitLearnInterface"
+        model.package_name == "MLJScikitLearnInterface" ||
+        # can be removed after resolution of
+        # https://github.com/JuliaAI/FeatureSelection.jl/issues/15
+        # and a Model Registry update
+        model.name == "RecursiveFeatureElimination"
 end
 
 WITHOUT_DATASETS = vcat(WITHOUT_DATASETS, PATHOLOGIES)