Added support for the 'RandomizedSearchCV' model type

README.md

@@ -95,6 +95,7 @@ Java library and command-line application for converting [Scikit-Learn](https://
     * [`linear_model.TheilSenRegressor`](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.TheilSenRegressor.html)
   * Model Selection:
     * [`model_selection.GridSearchCV`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html)
+    * [`model_selection.RandomizedSearchCV`](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.RandomizedSearchCV.html)
   * Multiclass classification:
     * [`multiclass.OneVsRestClassifier`](https://scikit-learn.org/stable/modules/generated/sklearn.multiclass.OneVsRestClassifier.html)
   * Naive Bayes:

src/main/java/sklearn/model_selection/GridSearchCV.java → src/main/java/sklearn/model_selection/EstimatorSearcher.java

@@ -22,9 +22,9 @@
 import org.jpmml.sklearn.PyClassDict;
 import sklearn.Estimator;
 
-public class GridSearchCV extends PyClassDict implements Castable {
+public class EstimatorSearcher extends PyClassDict implements Castable {
 
-	public GridSearchCV(String module, String name){
+	public EstimatorSearcher(String module, String name){
 		super(module, name);
 	}
 

src/main/resources/META-INF/sklearn2pmml.properties

@@ -98,7 +98,8 @@ sklearn.linear_model.(_sqd_fast|sgd_fast).SquaredHinge = sklearn.linear_model.st
 sklearn.linear_model.(_stochastic_gradient|stochastic_gradient).SGDClassifier = sklearn.linear_model.stochastic_gradient.SGDClassifier
 sklearn.linear_model.(_stochastic_gradient|stochastic_gradient).SGDRegressor = sklearn.linear_model.LinearRegressor
 sklearn.linear_model.(_theil_sen|theil_sen).TheilSenRegressor = sklearn.linear_model.LinearRegressor
-sklearn.model_selection._search.GridSearchCV = sklearn.model_selection.GridSearchCV
+sklearn.model_selection._search.GridSearchCV = sklearn.model_selection.EstimatorSearcher
+sklearn.model_selection._search.RandomizedSearchCV = sklearn.model_selection.EstimatorSearcher
 sklearn.multiclass.OneVsRestClassifier =
 sklearn.naive_bayes.GaussianNB =
 sklearn.neighbors.(_classification|classification).KNeighborsClassifier = sklearn.neighbors.KNeighborsClassifier

src/test/resources/csv/NuSVCVersicolor.csv

@@ -100,7 +100,7 @@ Species
 1
 1
 0
-1
+0
 0
 0
 0
@@ -120,9 +120,9 @@ Species
 0
 1
 0
-1
 0
-1
+0
+0
 0
 0
 1
@@ -141,7 +141,7 @@ Species
 0
 0
 0
-1
+0
 0
 0
 0

src/test/resources/main.py

@@ -19,7 +19,7 @@
 from sklearn.impute import MissingIndicator, SimpleImputer
 from sklearn.isotonic import IsotonicRegression
 from sklearn.linear_model import ARDRegression, BayesianRidge, ElasticNet, ElasticNetCV, HuberRegressor, LarsCV, Lasso, LassoCV, LassoLarsCV, LinearRegression, LogisticRegression, LogisticRegressionCV, OrthogonalMatchingPursuitCV, Ridge, RidgeCV, RidgeClassifier, RidgeClassifierCV, SGDClassifier, SGDRegressor, TheilSenRegressor
-from sklearn.model_selection import GridSearchCV
+from sklearn.model_selection import GridSearchCV, RandomizedSearchCV
 from sklearn.multiclass import OneVsRestClassifier
 from sklearn.naive_bayes import GaussianNB
 from sklearn.neighbors import KNeighborsClassifier, KNeighborsRegressor
@@ -344,7 +344,7 @@ def build_versicolor(classifier, name, with_proba = True, **pmml_options):
 	build_versicolor(SGDClassifier(max_iter = 100, random_state = 13), "SGDVersicolor", with_proba = False)
 	build_versicolor(SGDClassifier(loss = "log", max_iter = 100, random_state = 13), "SGDLogVersicolor")
 	build_versicolor(GridSearchCV(SVC(gamma = "auto"), {"C" : [1, 3, 5]}), "SVCVersicolor", with_proba = False)
-	build_versicolor(NuSVC(gamma = "auto"), "NuSVCVersicolor", with_proba = False)
+	build_versicolor(RandomizedSearchCV(NuSVC(gamma = "auto"), {"nu" : [0.3, 0.4, 0.5, 0.6]}), "NuSVCVersicolor", with_proba = False)
 
 versicolor_X, versicolor_y = load_versicolor("Versicolor")