Updated util dependencies.

Author: Olcay Taner YILDIZ

Test/C45Test.cpp

@@ -39,22 +39,6 @@ TEST_CASE("C45Test-testTrain") {
     }
     dataDefinition = DataDefinition(attributeTypes);
     tictactoe = DataSet(dataDefinition, ",", "tictactoe.data");
-    attributeTypes.clear();
-    for (int i = 0; i < 8; i++) {
-        attributeTypes.emplace_back(AttributeType::DISCRETE);
-    }
-    dataDefinition = DataDefinition(attributeTypes);
-    nursery = DataSet(dataDefinition, ",", "nursery.data");
-    attributeTypes.clear();
-    for (int i = 0; i < 6; i++) {
-        if (i % 2 == 0) {
-            attributeTypes.emplace_back(AttributeType::DISCRETE);
-        } else {
-            attributeTypes.emplace_back(AttributeType::CONTINUOUS);
-        }
-    }
-    dataDefinition = DataDefinition(attributeTypes);
-    chess = DataSet(dataDefinition, ",", "chess.data");
     C45 c45 = C45();
     C45Parameter* c45Parameter = new C45Parameter(1, true, 0.2);
     InstanceList instanceList = iris.getInstanceList();
@@ -72,10 +56,4 @@ TEST_CASE("C45Test-testTrain") {
     instanceList = tictactoe.getInstanceList();
     c45.train(instanceList, c45Parameter);
     REQUIRE_THAT(3.24, Catch::Matchers::WithinAbs(100 * c45.test(tictactoe.getInstanceList())->getErrorRate(), 0.01));
-    instanceList = nursery.getInstanceList();
-    c45.train(instanceList, c45Parameter);
-    REQUIRE_THAT(0.31, Catch::Matchers::WithinAbs(100 * c45.test(nursery.getInstanceList())->getErrorRate(), 0.01));
-    instanceList = chess.getInstanceList();
-    c45.train(instanceList, c45Parameter);
-    REQUIRE_THAT(47.59, Catch::Matchers::WithinAbs(100 * c45.test(chess.getInstanceList())->getErrorRate(), 0.01));
 }

Test/RandomForestTest.cpp

@@ -39,12 +39,6 @@ TEST_CASE("RandomForestTest-testTrain") {
     }
     dataDefinition = DataDefinition(attributeTypes);
     tictactoe = DataSet(dataDefinition, ",", "tictactoe.data");
-    attributeTypes.clear();
-    for (int i = 0; i < 8; i++) {
-        attributeTypes.emplace_back(AttributeType::DISCRETE);
-    }
-    dataDefinition = DataDefinition(attributeTypes);
-    nursery = DataSet(dataDefinition, ",", "nursery.data");
     RandomForest randomForest = RandomForest();
     auto* randomForestParameter = new RandomForestParameter(1, 100, 35);
     InstanceList instanceList = iris.getInstanceList();
@@ -62,7 +56,4 @@ TEST_CASE("RandomForestTest-testTrain") {
     instanceList = tictactoe.getInstanceList();
     randomForest.train(instanceList, randomForestParameter);
     REQUIRE_THAT(0.0, Catch::Matchers::WithinAbs(100 * randomForest.test(tictactoe.getInstanceList())->getErrorRate(), 0.01));
-    instanceList = nursery.getInstanceList();
-    randomForest.train(instanceList, randomForestParameter);
-    REQUIRE_THAT(0.0, Catch::Matchers::WithinAbs(100 * randomForest.test(nursery.getInstanceList())->getErrorRate(), 0.01));
 }

cmake-build-debug/lib/libClassification.a

@@ -4,7 +4,7 @@
 class ClassificationConan(ConanFile):
     name = "Classification"
     version = "1.0.0"
-    requires = ["Dictionary/1.0.0", "Sampling/1.0.0", "DataStructure/1.0.0"]
+    requires = ["Dictionary/1.0.0", "Sampling/1.0.0", "DataStructure/1.0.0", "Util/1.0.0"]
     license = "GPL License"
     author = "Olcay Taner Yildiz olcay.yildiz@ozyegin.edu.tr"
     url = "https://github.com/StarlangSoftware/Classification-CPP"

conanfile.py

@@ -2,6 +2,7 @@
 Dictionary/1.0.0
 Sampling/1.0.0
 DataStructure/1.0.0
+Util/1.0.0
 
 [generators]
 cmake

conanfile.txt

@@ -5,12 +5,9 @@
 #include <fstream>
 #include "DataSet.h"
 #include "Dictionary/Word.h"
-#include "../InstanceList/Partition.h"
 #include "../Instance/CompositeInstance.h"
 #include "../Attribute/ContinuousAttribute.h"
 #include "../Attribute/DiscreteAttribute.h"
-#include "../Attribute/BinaryAttribute.h"
-#include "../Attribute/DiscreteIndexedAttribute.h"
 
 /**
  * Constructor for generating a new {@link DataSet}.

src/DataSet/DataSet.cpp

@@ -31,8 +31,8 @@ void KFoldRun::runExperiment(Classifier *classifier, Parameter *parameter, Exper
  * @return An array of performances: result. result[i] is the performance of the classifier on the i'th fold.
  */
 ExperimentPerformance *KFoldRun::execute(Experiment experiment) {
-    ExperimentPerformance* result = new ExperimentPerformance();
-    KFoldCrossValidation<Instance*>* crossValidation = new KFoldCrossValidation<Instance*>(experiment.getDataSet().getInstances(), K, experiment.getParameter()->getSeed());
+    auto* result = new ExperimentPerformance();
+    auto* crossValidation = new KFoldCrossValidation<Instance*>(experiment.getDataSet().getInstances(), K, experiment.getParameter()->getSeed());
     runExperiment(experiment.getClassifier(), experiment.getParameter(), result, crossValidation);
     return result;
 }

src/Experiment/KFoldRun.cpp

@@ -4,7 +4,6 @@
 
 #include <StratifiedKFoldCrossValidation.h>
 #include "StratifiedMxKFoldRunSeparateTest.h"
-#include "../InstanceList/Partition.h"
 
 /**
  * Constructor for StratifiedMxKFoldRunSeparateTest class. Basically sets K parameter of the K-fold cross-validation and M for the number of times.

src/Experiment/StratifiedMxKFoldRunSeparateTest.cpp

@@ -9,6 +9,7 @@
 #include "../Attribute/DiscreteAttribute.h"
 #include "../Attribute/DiscreteIndexedAttribute.h"
 #include "../Attribute/ContinuousAttribute.h"
+#include "RandomArray.h"
 
 using namespace std;
 /**
@@ -90,12 +91,7 @@ Partition::Partition(InstanceList &list, double ratio, int seed, bool stratified
         add(new InstanceList());
         add(new InstanceList());
         distribution = list.classDistribution();
-        vector<int> randomArray;
-        randomArray.reserve(list.size());
-        for (int i = 0; i < list.size(); i++){
-            randomArray.push_back(i);
-        }
-        std::shuffle(randomArray.begin(), randomArray.end(), default_random_engine(seed));
+        vector<int> randomArray = RandomArray::indexArray(list.size(), seed);
         for (int i = 0; i < list.size(); i++) {
             Instance* instance = list.get(randomArray.at(i));
             if (counts.count(instance->getClassLabel()) < list.size() * ratio * distribution.getProbability(instance->getClassLabel())) {

src/InstanceList/Partition.cpp

@@ -11,6 +11,7 @@
 #include "../../Attribute/ContinuousAttribute.h"
 #include "../../Classifier/Classifier.h"
 #include "../../Instance/CompositeInstance.h"
+#include "RandomArray.h"
 
 /**
  * The entropyForDiscreteAttribute method takes an attributeIndex and creates an ArrayList of DiscreteDistribution.
@@ -124,14 +125,14 @@ DecisionNode::DecisionNode(InstanceList data, DecisionCondition condition, Rando
         return;
     }
     vector<int> indexList;
-    indexList.reserve(data.get(0)->attributeSize());
-    for (int i = 0; i < data.get(0)->attributeSize(); i++) {
-        indexList.push_back(i);
-    }
     if (parameter != nullptr && parameter->getAttributeSubsetSize() < data.get(0)->attributeSize()) {
-        std::shuffle(indexList.begin(), indexList.end(), default_random_engine(parameter->getSeed()));
+        indexList = RandomArray::indexArray(data.get(0)->attributeSize(), parameter->getSeed());
         size = parameter->getAttributeSubsetSize();
     } else {
+        indexList.reserve(data.get(0)->attributeSize());
+        for (int i = 0; i < data.get(0)->attributeSize(); i++) {
+            indexList.push_back(i);
+        }
         size = data.get(0)->attributeSize();
     }
     classDistribution = data.classDistribution();