Add XGBoost Classifier

src/PyClassifiers/XGBoost.cc

@@ -1,16 +1,11 @@
 #include "XGBoost.h"
 
-
-
-
 //See https ://stackoverflow.com/questions/36071672/using-xgboost-in-c
-
-
-
-
-
-
 namespace pywrap {
+    XGBoost::XGBoost() : PyClassifier("xgboost", "XGBClassifier")
+    {
+        validHyperparameters = { "tree_method", "early_stopping_rounds", "n_jobs" };
+    }
     std::string XGBoost::version()
     {
         return callMethodString("1.0");

src/PyClassifiers/XGBoost.h

@@ -5,7 +5,7 @@
 namespace pywrap {
     class XGBoost : public PyClassifier {
     public:
-        XGBoost() : PyClassifier("xgboost", "XGBClassifier") {};
+        XGBoost();
         ~XGBoost() = default;
         std::string version();
     };

tests/TestPythonClassifiers.cc

@@ -8,8 +8,10 @@
 #include "STree.h"
 #include "SVC.h"
 #include "RandomForest.h"
+#include "XGBoost.h"
 #include "ODTE.h"
 #include "TestUtils.h"
+#include <nlohmann/json.hpp>
 
 TEST_CASE("Test Python Classifiers score", "[PyClassifiers]")
 {
@@ -72,3 +74,13 @@ TEST_CASE("Get num features & num edges", "[PyClassifiers]")
     REQUIRE(clf.getNumberOfNodes() == 10);
     REQUIRE(clf.getNumberOfEdges() == 10);
 }
+TEST_CASE("XGBoost", "[PyClassifiers]")
+{
+    auto raw = RawDatasets("iris", true);
+    auto clf = pywrap::XGBoost();
+    clf.fit(raw.Xt, raw.yt, raw.featurest, raw.classNamet, raw.statest);
+    nlohmann::json hyperparameters = { "n_jobs=1" };
+    clf.setHyperparameters(hyperparameters);
+    auto score = clf.score(raw.Xt, raw.yt);
+    REQUIRE(score == Catch::Approx(0.98).epsilon(raw.epsilon));
+}