Relintai/pmlpp
1
0
Fork 0
mirror of https://github.com/Relintai/pmlpp.git synced 2025-01-18 15:07:16 +01:00
Code Issues Packages Projects Releases Wiki Activity

Cleaned up OutlierFinder.

Browse Source
This commit is contained in:
Relintai 2023-02-09 15:30:33 +01:00
parent 8f7177aaac
commit d341f6f8d0
6 changed files with 157 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

137
mlpp/outlier_finder/outlier_finder.cpp
View File

@ -5,37 +5,134 @@
//
#include "outlier_finder.h"
#include "../stat/stat.h"
#include <iostream>
MLPPOutlierFinder::MLPPOutlierFinder(int threshold) :
threshold(threshold) {
real_t MLPPOutlierFinder::get_threshold() {
return _threshold;
}
void MLPPOutlierFinder::set_threshold(real_t val) {
_threshold = val;
}
std::vector<std::vector<real_t>> MLPPOutlierFinder::modelSetTest(std::vector<std::vector<real_t>> inputSet) {
MLPPStat stat;
std::vector<std::vector<real_t>> outliers;
outliers.resize(inputSet.size());
for (int i = 0; i < inputSet.size(); i++) {
for (int j = 0; j < inputSet[i].size(); j++) {
real_t z = (inputSet[i][j] - stat.mean(inputSet[i])) / stat.standardDeviation(inputSet[i]);
if (abs(z) > threshold) {
outliers[i].push_back(inputSet[i][j]);
Vector<Vector<real_t>> MLPPOutlierFinder::model_set_test(const Ref<MLPPMatrix> &input_set) {
ERR_FAIL_COND_V(!input_set.is_valid(), Vector<Vector<real_t>>());
MLPPStat stat;
Size2i input_set_size = input_set->size();
Vector<Vector<real_t>> outliers;
outliers.resize(input_set_size.y);
Ref<MLPPVector> input_set_i_row_tmp;
input_set_i_row_tmp.instance();
input_set_i_row_tmp->resize(input_set_size.x);
for (int i = 0; i < input_set_size.y; ++i) {
input_set->get_row_into_mlpp_vector(i, input_set_i_row_tmp);
real_t meanv = stat.meanv(input_set_i_row_tmp);
real_t s_dev_v = stat.standard_deviationv(input_set_i_row_tmp);
for (int j = 0; j < input_set_size.x; ++j) {
real_t input_set_i_j = input_set->get_element(i, j);
real_t z = (input_set_i_j - meanv) / s_dev_v;
if (ABS(z) > _threshold) {
outliers.write[i].push_back(input_set_i_j);
}
}
}
return outliers;
}
std::vector<real_t> MLPPOutlierFinder::modelTest(std::vector<real_t> inputSet) {
MLPPStat stat;
std::vector<real_t> outliers;
for (int i = 0; i < inputSet.size(); i++) {
real_t z = (inputSet[i] - stat.mean(inputSet)) / stat.standardDeviation(inputSet);
if (abs(z) > threshold) {
outliers.push_back(inputSet[i]);
Array MLPPOutlierFinder::model_set_test_bind(const Ref<MLPPMatrix> &input_set) {
Vector<Vector<real_t>> res = model_set_test(input_set);
Array arr;
for (int i = 0; i < res.size(); ++i) {
//will get converted to PoolRealArray
arr.push_back(Variant(res[i]));
}
return arr;
}
PoolVector2iArray MLPPOutlierFinder::model_set_test_indices(const Ref<MLPPMatrix> &input_set) {
ERR_FAIL_COND_V(!input_set.is_valid(), PoolVector2iArray());
MLPPStat stat;
Size2i input_set_size = input_set->size();
PoolVector2iArray outliers;
Ref<MLPPVector> input_set_i_row_tmp;
input_set_i_row_tmp.instance();
input_set_i_row_tmp->resize(input_set_size.x);
for (int i = 0; i < input_set_size.y; ++i) {
input_set->get_row_into_mlpp_vector(i, input_set_i_row_tmp);
real_t meanv = stat.meanv(input_set_i_row_tmp);
real_t s_dev_v = stat.standard_deviationv(input_set_i_row_tmp);
for (int j = 0; j < input_set_size.x; ++j) {
real_t z = (input_set->get_element(i, j) - meanv) / s_dev_v;
if (ABS(z) > _threshold) {
outliers.push_back(Vector2i(j, i));
}
}
}
return outliers;
}
PoolRealArray MLPPOutlierFinder::model_test(const Ref<MLPPVector> &input_set) {
ERR_FAIL_COND_V(!input_set.is_valid(), PoolRealArray());
MLPPStat stat;
PoolRealArray outliers;
real_t mean = stat.meanv(input_set);
real_t s_dev = stat.standard_deviationv(input_set);
int input_set_size = input_set->size();
const real_t *input_set_ptr = input_set->ptr();
for (int i = 0; i < input_set_size; ++i) {
real_t input_set_i = input_set_ptr[i];
real_t z = (input_set_i - mean) / s_dev;
if (ABS(z) > _threshold) {
outliers.push_back(input_set_i);
}
}
return outliers;
}
MLPPOutlierFinder::MLPPOutlierFinder(real_t threshold) {
_threshold = threshold;
}
MLPPOutlierFinder::MLPPOutlierFinder() {
_threshold = 0;
}
MLPPOutlierFinder::~MLPPOutlierFinder() {
}
void MLPPOutlierFinder::_bind_methods() {
ClassDB::bind_method(D_METHOD("get_threshold"), &MLPPOutlierFinder::get_threshold);
ClassDB::bind_method(D_METHOD("set_threshold", "val"), &MLPPOutlierFinder::set_threshold);
ADD_PROPERTY(PropertyInfo(Variant::REAL, "threshold"), "set_threshold", "get_threshold");
ClassDB::bind_method(D_METHOD("model_set_test", "input_set"), &MLPPOutlierFinder::model_set_test_bind);
ClassDB::bind_method(D_METHOD("model_set_test_indices", "input_set"), &MLPPOutlierFinder::model_set_test_indices);
ClassDB::bind_method(D_METHOD("model_test", "input_set"), &MLPPOutlierFinder::model_test);
}

32
mlpp/outlier_finder/outlier_finder.h
View File

@ -10,20 +10,34 @@
#include "core/math/math_defs.h"
#include <vector>
#include "core/object/reference.h"
#include "../lin_alg/mlpp_matrix.h"
#include "../lin_alg/mlpp_vector.h"
class MLPPOutlierFinder : public Reference {
GDCLASS(MLPPOutlierFinder, Reference);
class MLPPOutlierFinder {
public:
// Cnstr
MLPPOutlierFinder(int threshold);
real_t get_threshold();
void set_threshold(real_t val);
std::vector<std::vector<real_t>> modelSetTest(std::vector<std::vector<real_t>> inputSet);
std::vector<real_t> modelTest(std::vector<real_t> inputSet);
Vector<Vector<real_t>> model_set_test(const Ref<MLPPMatrix> &input_set);
Array model_set_test_bind(const Ref<MLPPMatrix> &input_set);
// Variables required
int threshold;
PoolVector2iArray model_set_test_indices(const Ref<MLPPMatrix> &input_set);
PoolRealArray model_test(const Ref<MLPPVector> &input_set);
MLPPOutlierFinder(real_t threshold);
MLPPOutlierFinder();
~MLPPOutlierFinder();
protected:
static void _bind_methods();
real_t _threshold;
};
#endif /* OutlierFinder_hpp */

4
mlpp/stat/stat.cpp
View File

@ -133,6 +133,10 @@ real_t MLPPStat::meanv(const Ref<MLPPVector> &x) {
return sum / x_size;
}
real_t MLPPStat::standard_deviationv(const Ref<MLPPVector> &x) {
return Math::sqrt(variancev(x));
}
real_t MLPPStat::variancev(const Ref<MLPPVector> &x) {
real_t x_mean = meanv(x);

1
mlpp/stat/stat.h
View File

@ -39,6 +39,7 @@ public:
real_t chebyshevIneq(const real_t k);
real_t meanv(const Ref<MLPPVector> &x);
real_t standard_deviationv(const Ref<MLPPVector> &x);
real_t variancev(const Ref<MLPPVector> &x);
real_t covariancev(const Ref<MLPPVector> &x, const Ref<MLPPVector> &y);

2
register_types.cpp
View File

@ -38,6 +38,7 @@ SOFTWARE.
#include "mlpp/kmeans/kmeans.h"
#include "mlpp/knn/knn.h"
#include "mlpp/outlier_finder/outlier_finder.h"
#include "mlpp/pca/pca.h"
#include "mlpp/uni_lin_reg/uni_lin_reg.h"
#include "mlpp/wgan/wgan.h"
@ -67,6 +68,7 @@ void register_pmlpp_types(ModuleRegistrationLevel p_level) {
ClassDB::register_class<MLPPWGAN>();
ClassDB::register_class<MLPPPCA>();
ClassDB::register_class<MLPPUniLinReg>();
ClassDB::register_class<MLPPOutlierFinder>();
ClassDB::register_class<MLPPDataESimple>();
ClassDB::register_class<MLPPDataSimple>();

12
test/mlpp_tests.cpp
View File

@ -48,6 +48,7 @@
#include "../mlpp/wgan/wgan.h"
#include "../mlpp/mlp/mlp_old.h"
#include "../mlpp/outlier_finder/outlier_finder_old.h"
#include "../mlpp/pca/pca_old.h"
#include "../mlpp/uni_lin_reg/uni_lin_reg_old.h"
#include "../mlpp/wgan/wgan_old.h"
@ -855,8 +856,15 @@ void MLPPTests::test_outlier_finder(bool ui) {
// Outlier Finder
//std::vector<real_t> inputSet = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 23554332523523 };
std::vector<real_t> inputSet = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 23554332 };
MLPPOutlierFinder outlierFinder(2); // Any datapoint outside of 2 stds from the mean is marked as an outlier.
alg.printVector(outlierFinder.modelTest(inputSet));
MLPPOutlierFinderOld outlierFinderOld(2); // Any datapoint outside of 2 stds from the mean is marked as an outlier.
alg.printVector(outlierFinderOld.modelTest(inputSet));
Ref<MLPPVector> input_set;
input_set.instance();
input_set->set_from_std_vector(inputSet);
MLPPOutlierFinder outlier_finder(2); // Any datapoint outside of 2 stds from the mean is marked as an outlier.
PLOG_MSG(Variant(outlier_finder.model_test(input_set)));
}
void MLPPTests::test_new_math_functions() {
MLPPLinAlg alg;