apaam/mlpack
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

optimize binarize and add binarize executable

Browse Source
This commit is contained in:
Keon Kim
2016-06-14 21:11:15 +09:00
parent d797162a83
commit cd0a3770a1
5 changed files with 131 additions and 113 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/mlpack/core/data/binarize.hpp
+19 -81
View File
@@ -9,41 +9,10 @@
#define MLPACK_CORE_DATA_BINARIZE_HPP
#include <mlpack/core.hpp>
#include <omp.h>
namespace mlpack {
namespace data {
/**
* Given an input dataset and threshold, set values greater than threshold to
* 1 and values less than or equal to the threshold to 0. This overload takes
* a dimension and applys the changes to the given dimension.
*
* @code
* arma::mat input = loadData();
* double threshold = 0;
* size_t dimension = 0;
*
* // Binarize the first dimension. All positive values in the first dimension
* // will be set to 1 and the values less than or equal to 0 will become 0.
* Binarize(input, threshold, dimension);
* @endcode
*
* @param input Input matrix to Binarize.
* @param threshold Threshold can by any number.
* @param dimension Feature to apply the Binarize function.
*/
template<typename T>
void Binarize(arma::Mat<T>& input,
const double threshold,
const size_t dimension)
{
for (size_t i = 0; i < input.n_cols; ++i)
{
if (input(dimension, i) > threshold)
input(dimension, i) = 1;
else
input(dimension, i) = 0;
}
}
/**
* Given an input dataset and threshold, set values greater than threshold to
@@ -51,46 +20,13 @@ void Binarize(arma::Mat<T>& input,
* the changes to all dimensions.
*
* @code
* arma::mat input = loadData();
* double threshold = 0;
*
* // Binarize the whole Matrix. All positive values in will be set to 1 and
* // the values less than or equal to 0 will become 0.
* Binarize(input, threshold);
* @endcode
*
* @param input Input matrix to Binarize.
* @param threshold Threshold can by any number.
*/
template<typename T>
void Binarize(arma::Mat<T>& input,
const double threshold)
{
for (size_t i = 0; i < input.n_cols; ++i)
{
for (size_t j = 0; j < input.n_rows; ++j)
{
if (input(i, j) > threshold)
input(i, j) = 1;
else
input(i, j) = 0;
}
}
}
/**
* Given an input dataset and threshold, set values greater than threshold to
* 1 and values less than or equal to the threshold to 0. This overload applies
* the changes to all dimensions.
*
* @code
* arma::mat input = loadData();
* arma::mat output;
* arma::Mat<double> input = loadData();
* arma::Mat<double> output;
* double threshold = 0.5;
*
* // Binarize the whole Matrix. All positive values in will be set to 1 and
* // the values less than or equal to 0.5 will become 0.
* Binarize(input, output, threshold);
* Binarize<double>(input, output, threshold);
* @endcode
*
* @param input Input matrix to Binarize.
@@ -104,15 +40,17 @@ void Binarize(const arma::Mat<T>& input,
{
output.copy_size(input);
for (size_t i = 0; i < input.n_cols; ++i)
const size_t totalElems = static_cast<size_t>(input.n_elem);
const T *inPtr = input.memptr();
T *outPtr = output.memptr();
#pragma omp parallel for
for (size_t i = 0; i < totalElems; ++i)
{
for (size_t j = 0; j < input.n_rows; ++j)
{
if (input(i, j) > threshold)
output(i, j) = 1;
else
output(i, j) = 0;
}
if (inPtr[i] < threshold)
outPtr[i] = 0;
else
outPtr[i] = 1;
}
}
@@ -122,14 +60,14 @@ void Binarize(const arma::Mat<T>& input,
* a dimension and applys the changes to the given dimension.
*
* @code
* arma::mat input = loadData();
* arma::mat output;
* arma::Mat<double> input = loadData();
* arma::Mat<double> output;
* double threshold = 0.5;
* size_t dimension = 0;
*
* // Binarize the first dimension. All positive values in the first dimension
* // will be set to 1 and the values less than or equal to 0 will become 0.
* Binarize(input, output, threshold, dimension);
* Binarize<double>(input, output, threshold, dimension);
* @endcode
*
* @param input Input matrix to Binarize.
@@ -137,15 +75,15 @@ void Binarize(const arma::Mat<T>& input,
* @param threshold Threshold can by any number.
* @param dimension Feature to apply the Binarize function.
*/
template<typename T>
void Binarize(const arma::Mat<T>& input,
arma::Mat<T>& output,
const double threshold,
const size_t dimension)
{
output(input);
output = input;
#pragma omp parallel for
for (size_t i = 0; i < input.n_cols; ++i)
{
if (input(dimension, i) > threshold)
src/mlpack/methods/preprocess/CMakeLists.txt
+1
View File
@@ -14,5 +14,6 @@ set(MLPACK_SRCS ${MLPACK_SRCS} ${DIR_SRCS} PARENT_SCOPE)
#add_cli_executable(preprocess_stats)
add_cli_executable(preprocess_split)
add_cli_executable(preprocess_binarize)
#add_cli_executable(preprocess_scan)
#add_cli_executable(preprocess_imputer)
src/mlpack/methods/preprocess/preprocess_binarize_main.cpp
+71
View File
@@ -0,0 +1,71 @@
/**
* @file preprocess_binarize_main.cpp
* @author Keon Kim
*
* split data CLI executable
*/
#include <mlpack/core.hpp>
#include <mlpack/core/data/binarize.hpp>
PROGRAM_INFO("Split Data", "This utility takes a dataset and optionally labels "
"and splits ");
// Define parameters for data.
PARAM_STRING_REQ("input_file", "File containing data,", "i");
// Define optional parameters.
PARAM_STRING("output_file", "File to save the output,", "o", "");
PARAM_INT("feature", "File containing labels", "f", 0);
PARAM_DOUBLE("threshold", "Ratio of test set, if not set,"
"the threshold defaults to 0.0", "t", 0.0);
using namespace mlpack;
using namespace arma;
using namespace std;
int main(int argc, char** argv)
{
// Parse command line options.
CLI::ParseCommandLine(argc, argv);
const string inputFile = CLI::GetParam<string>("input_file");
const string outputFile = CLI::GetParam<string>("output_file");
const size_t feature = (size_t) CLI::GetParam<int>("feature");
const double threshold = CLI::GetParam<double>("threshold");
// Check on data parameters.
if (!CLI::HasParam("feature"))
Log::Warn << "You did not specify --feature, so the program will perform "
<< "binarize on every features." << endl;
if (!CLI::HasParam("threshold"))
Log::Warn << "You did not specify --threshold, so the threhold "
<< "will be automatically set to '0.0'." << endl;
if (!CLI::HasParam("output_file"))
Log::Warn << "You did not specify --output_file, so no result will be"
<< "saved." << endl;
// Load the data.
arma::mat input;
arma::mat output;
data::Load(inputFile, input, true);
Timer::Start("binarize");
if (CLI::HasParam("feature"))
{
data::Binarize<double>(input, output, threshold, feature);
}
else
{
// binarize the whole data
data::Binarize<double>(input, output, threshold);
}
Timer::Stop("binarize");
Log::Info << "input" << endl;
Log::Info << input << endl;
Log::Info << "output" << endl;
Log::Info << output << endl;
if (CLI::HasParam("output_file"))
data::Save(outputFile, output, false);
}
src/mlpack/methods/preprocess/preprocess_split_main.cpp
+2 -2
View File
@@ -1,6 +1,6 @@
/**
* @file preprocess_split_main.cpp
* @author Keon Woo Kim
* @author Keon Kim
*
* split data CLI executable
*/
@@ -78,7 +78,7 @@ int main(int argc, char** argv)
{
trainingFile = "train_" + inputFile;
Log::Warn << "You did not specify --training_file, so the training set file"
<< " name will be automatically set to '" << trainingFile << "'."
<< " name will be automatically set to '" << trainingFile << "'."
<< endl;
}
if (testFile.empty())
src/mlpack/tests/binarize_test.cpp
+38 -30
View File
@@ -17,42 +17,50 @@ using namespace mlpack::data;
BOOST_AUTO_TEST_SUITE(BinarizeTest);
/**
* Compare the binarized data with answer.
*
* @param input The original data set before Binarize.
* @param answer The data want to compare with the input.
*/
void CheckAnswer(const mat& input,
const umat& answer)
BOOST_AUTO_TEST_CASE(BinerizeOneDimension)
{
for (size_t i = 0; i < input.n_cols; ++i)
{
const mat& lhsCol = input.col(i);
const umat& rhsCol = answer.col(i);
for (size_t j = 0; j < lhsCol.n_rows; ++j)
{
if (std::abs(rhsCol(j)) < 1e-5)
BOOST_REQUIRE_SMALL(lhsCol(j), 1e-5);
else
BOOST_REQUIRE_CLOSE(lhsCol(j), rhsCol(j), 1e-5);
}
}
mat input;
input << 1 << 2 << 3 << endr
<< 4 << 5 << 6 << endr // this row will be tested
<< 7 << 8 << 9;
mat output;
const double threshold = 5.0;
const size_t dimension = 1;
Binarize<double>(input, output, threshold, dimension);
BOOST_REQUIRE_CLOSE(input(0, 0), 1, 1e-5); // 1
BOOST_REQUIRE_CLOSE(input(0, 1), 2, 1e-5); // 2
BOOST_REQUIRE_CLOSE(input(0, 2), 3, 1e-5); // 3
BOOST_REQUIRE_SMALL(input(1, 0), 1e-5); // 4 target
BOOST_REQUIRE_SMALL(input(1, 1), 1e-5); // 5 target
BOOST_REQUIRE_CLOSE(input(1, 2), 1, 1e-5); // 6 target
BOOST_REQUIRE_CLOSE(input(2, 0), 7, 1e-5); // 7
BOOST_REQUIRE_CLOSE(input(2, 1), 8, 1e-5); // 8
BOOST_REQUIRE_CLOSE(input(2, 2), 9, 1e-5); // 9
}
BOOST_AUTO_TEST_CASE(BinarizeThreshold)
BOOST_AUTO_TEST_CASE(BinerizeAll)
{
mat input(10, 10, fill::randu); // fill input with randome Number
mat constMat(10, 10);
double threshold = math::Random(); // random number threshold
constMat.fill(threshold);
mat input;
input << 1 << 2 << 3 << endr
<< 4 << 5 << 6 << endr // this row will be tested
<< 7 << 8 << 9;
umat answer = input > constMat;
mat output;
const double threshold = 5.0;
const size_t dimension = 1;
Binarize<double>(input, output, threshold);
// Binarize every values inside the matrix with threshold of 0;
Binarize(input, threshold);
CheckAnswer(input, answer);
BOOST_REQUIRE_SMALL(input(0, 0), 1e-5); // 1
BOOST_REQUIRE_SMALL(input(0, 1), 1e-5); // 2
BOOST_REQUIRE_SMALL(input(0, 2), 1e-5); // 3
BOOST_REQUIRE_SMALL(input(1, 0), 1e-5); // 4
BOOST_REQUIRE_SMALL(input(1, 1), 1e-5); // 5
BOOST_REQUIRE_CLOSE(input(1, 2), 1.0, 1e-5); // 6
BOOST_REQUIRE_CLOSE(input(2, 0), 1.0, 1e-5); // 7
BOOST_REQUIRE_CLOSE(input(2, 1), 1.0, 1e-5); // 8
BOOST_REQUIRE_CLOSE(input(2, 2), 1.0, 1e-5); // 9
}
BOOST_AUTO_TEST_SUITE_END();