BaggingMachine_unittest.cc

#include "machine/MockMachine.h"
#include "features/MockFeatures.h"
#include "labels/MockLabels.h"
#include <shogun/features/DenseFeatures.h>
#include <shogun/lib/SGMatrix.h>
#include <shogun/multiclass/tree/CARTree.h>
#include <shogun/lib/config.h>
#include <shogun/machine/BaggingMachine.h>
#include <shogun/evaluation/MulticlassAccuracy.h>
#include <shogun/ensemble/MajorityVote.h>
#include <gtest/gtest.h>

#define sunny 1.
#define overcast 2.
#define rain 3.

#define hot 1.
#define mild 2.
#define cool 3.

#define high 1.
#define normal 2.

#define weak 1.
#define strong 2.

using namespace shogun;
using ::testing::Return;

/** gmock REV 443 and freebsd doesn't play nicely */
#ifdef FREEBSD
TEST(BaggingMachine, DISABLED_mock_train)
#else
TEST(BaggingMachine, mock_train)
#endif
{
	using ::testing::NiceMock;
	using ::testing::_;
	using ::testing::InSequence;
	using ::testing::Mock;
	using ::testing::DefaultValue;

	int32_t bag_size = 20;
	int32_t num_bags = 10;

	NiceMock<MockCFeatures> features; features.ref();
	NiceMock<MockCLabels> labels; labels.ref();
	CBaggingMachine* bm = new CBaggingMachine(&features, &labels);
	NiceMock<MockCMachine> mm; mm.ref();
	CMajorityVote* mv = new CMajorityVote();

	bm->parallel->set_num_threads(1);
	bm->set_machine(&mm);
	bm->set_bag_size(bag_size);
	bm->set_num_bags(num_bags);
	bm->set_combination_rule(mv);

	ON_CALL(mm, train_machine(_))
		.WillByDefault(Return(true));

	ON_CALL(features, get_num_vectors())
		.WillByDefault(Return(100));

	{
		InSequence s;
		for (int i = 0; i < num_bags; i++) {
			EXPECT_CALL(mm, clone())
				.Times(1)
				.WillRepeatedly(Return(&mm));

			EXPECT_CALL(mm, train_machine(_))
				.Times(1)
				.WillRepeatedly(Return(true));

			mm.ref();
		}
	}

	bm->train();

	SG_UNREF(bm);
}

TEST(BaggingMachine,classify_CART)
{
	set_global_seed(1);
	SGMatrix<float64_t> data(4,14);

	//vector = [Outlook Temperature Humidity Wind]
	data(0,0)=sunny;
	data(1,0)=hot;
	data(2,0)=high;
	data(3,0)=weak;

	data(0,1)=sunny;
	data(1,1)=hot;
	data(2,1)=high;
	data(3,1)=strong;

	data(0,2)=overcast;
	data(1,2)=hot;
	data(2,2)=high;
	data(3,2)=weak;

	data(0,3)=rain;
	data(1,3)=mild;
	data(2,3)=high;
	data(3,3)=weak;

	data(0,4)=rain;
	data(1,4)=cool;
	data(2,4)=normal;
	data(3,4)=weak;

	data(0,5)=rain;
	data(1,5)=cool;
	data(2,5)=normal;
	data(3,5)=strong;

	data(0,6)=overcast;
	data(1,6)=cool;
	data(2,6)=normal;
	data(3,6)=strong;

	data(0,7)=sunny;
	data(1,7)=mild;
	data(2,7)=high;
	data(3,7)=weak;

	data(0,8)=sunny;
	data(1,8)=cool;
	data(2,8)=normal;
	data(3,8)=weak;

	data(0,9)=rain;
	data(1,9)=mild;
	data(2,9)=normal;
	data(3,9)=weak;

	data(0,10)=sunny;
	data(1,10)=mild;
	data(2,10)=normal;
	data(3,10)=strong;

	data(0,11)=overcast;
	data(1,11)=mild;
	data(2,11)=high;
	data(3,11)=strong;

	data(0,12)=overcast;
	data(1,12)=hot;
	data(2,12)=normal;
	data(3,12)=weak;

	data(0,13)=rain;
	data(1,13)=mild;
	data(2,13)=high;
	data(3,13)=strong;

	CDenseFeatures<float64_t>* feats=new CDenseFeatures<float64_t>(data);

	// yes 1. no 0.
	SGVector<float64_t> lab(14);
	lab[0]=0.0;
	lab[1]=0.0;
	lab[2]=1.0;
	lab[3]=1.0;
	lab[4]=1.0;
	lab[5]=0.0;
	lab[6]=1.0;
	lab[7]=0.0;
	lab[8]=1.0;
	lab[9]=1.0;
	lab[10]=1.0;
	lab[11]=1.0;
	lab[12]=1.0;
	lab[13]=0.0;

	SGVector<bool> ft=SGVector<bool>(4);
	ft[0]=true;
	ft[1]=true;
	ft[2]=true;
	ft[3]=true;

	CMulticlassLabels* labels=new CMulticlassLabels(lab);

	CCARTree* cart=new CCARTree();
	CMajorityVote* cv=new CMajorityVote();
	cart->set_feature_types(ft);
	CBaggingMachine* c=new CBaggingMachine(feats,labels);
	c->parallel->set_num_threads(1);
	c->set_machine(cart);
	c->set_bag_size(14);
	c->set_num_bags(10);
	c->set_combination_rule(cv);
	c->train(feats);

	SGMatrix<float64_t> test(4,5);
	test(0,0)=overcast;
	test(0,1)=rain;
	test(0,2)=sunny;
	test(0,3)=rain;
	test(0,4)=sunny;

	test(1,0)=hot;
	test(1,1)=cool;
	test(1,2)=mild;
	test(1,3)=mild;
	test(1,4)=hot;

	test(2,0)=normal;
	test(2,1)=high;
	test(2,2)=high;
	test(2,3)=normal;
	test(2,4)=normal;

	test(3,0)=strong;
	test(3,1)=strong;
	test(3,2)=weak;
	test(3,3)=weak;
	test(3,4)=strong;

	CDenseFeatures<float64_t>* test_feats=new CDenseFeatures<float64_t>(test);
	CMulticlassLabels* result=c->apply_multiclass(test_feats);
	SGVector<float64_t> res_vector=result->get_labels();

	EXPECT_EQ(1.0,res_vector[0]);
	EXPECT_EQ(0.0,res_vector[1]);
	EXPECT_EQ(0.0,res_vector[2]);
	EXPECT_EQ(1.0,res_vector[3]);
	EXPECT_EQ(0.0, res_vector[4]);

	CMulticlassAccuracy* eval=new CMulticlassAccuracy();
	EXPECT_NEAR(0.5, c->get_oob_error(eval), 1e-6);

	SG_UNREF(test_feats);
	SG_UNREF(result);
	SG_UNREF(c);
	SG_UNREF(eval);
}