diff --git a/Makefile b/Makefile
index 89f5c486793f56443c8d1edd2888939bdd8a36a3..7be3f52f04333a13dd3997d4c5684d3c2cb1b9ce 100644
--- a/Makefile
+++ b/Makefile
@@ -4,8 +4,11 @@ SPLFR_PATH=/usr/local/sicstus4.7.1/bin/splfr
all:
make -C src/methods/ada_boost splfr=$(SPLFR_PATH)
make -C src/methods/bayesian_linear_regression splfr=$(SPLFR_PATH)
+ make -C src/methods/cf splfr=$(SPLFR_PATH)
make -C src/methods/dbscan splfr=$(SPLFR_PATH)
make -C src/methods/emst splfr=$(SPLFR_PATH)
+ make -C src/methods/fastmks splfr=$(SPLFR_PATH)
+ make -C src/methods/kernel_pca splfr=$(SPLFR_PATH)
make -C src/methods/kmeans splfr=$(SPLFR_PATH)
make -C src/methods/lars splfr=$(SPLFR_PATH)
make -C src/methods/linear_regression splfr=$(SPLFR_PATH)
@@ -18,13 +21,15 @@ all:
make -C src/methods/random_forest splfr=$(SPLFR_PATH)
make -C src/methods/softmac_regression splfr=$(SPLFR_PATH)
make -C src/methods/approx_kfn splfr=$(SPLFR_PATH)
- make -C src/methods/fastmks splfr=$(SPLFR_PATH)
clean:
make -C src/methods/ada_boost clean
make -C src/methods/bayesian_linear_regression clean
+ make -C src/methods/cf clean
make -C src/methods/dbscan clean
make -C src/methods/emst clean
+ make -C src/methods/fastmks clean
+ make -C src/methods/kernel_pca clean
make -C src/methods/kmeans clean
make -C src/methods/lars clean
make -C src/methods/linear_regression clean
@@ -37,6 +42,5 @@ clean:
make -C src/methods/random_forest clean
make -C src/methods/softmac_regression clean
make -C src/methods/approx_kfn clean
- make -C src/methods/fastmks clean
\ No newline at end of file
diff --git a/src/methods/kernel_pca/Makefile b/src/methods/kernel_pca/Makefile
new file mode 100644
index 0000000000000000000000000000000000000000..76bab9734a8c694cee152b5f8fc902b41259c6e7
--- /dev/null
+++ b/src/methods/kernel_pca/Makefile
@@ -0,0 +1,8 @@
+splfr=/usr/local/sicstus4.7.1/bin/splfr
+
+METHOD_NAME=kernel_pca
+
+$(METHOD_NAME).so: $(METHOD_NAME).pl $(METHOD_NAME).cpp
+ $(splfr) -larmadillo -fopenmp -lmlpack -lstdc++ -cxx --struct $(METHOD_NAME).pl $(METHOD_NAME).cpp ../../helper_files/helper.cpp
+clean:
+ rm $(METHOD_NAME).so
diff --git a/src/methods/kernel_pca/kernel_pca.cpp b/src/methods/kernel_pca/kernel_pca.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..9dad7df4b073b2cf85d32a56da83b7903fa4086e
--- /dev/null
+++ b/src/methods/kernel_pca/kernel_pca.cpp
@@ -0,0 +1,189 @@
+#include <sicstus/sicstus.h>
+/* ex_glue.h is generated by splfr from the foreign/[2,3] facts.
+ Always include the glue header in your foreign resource code.
+*/
+#include "kernel_pca_glue.h"
+#include <mlpack/methods/kernel_pca/kernel_pca.hpp>
+#include <mlpack/core.hpp>
+
+#include <mlpack/methods/kernel_pca/kernel_rules/nystroem_method.hpp>
+
+#include <mlpack/methods/nystroem_method/ordered_selection.hpp>
+#include <mlpack/methods/nystroem_method/random_selection.hpp>
+
+// including helper functions for converting between arma structures and arrays
+#include "../../helper_files/helper.hpp"
+
+// some of the most used namespaces
+using namespace arma;
+using namespace mlpack;
+using namespace std;
+using namespace mlpack::kpca;
+using namespace mlpack::kernel;
+
+
+template<typename KernelType>
+void useNystroemKernelPCA(char const *nystroemMethod, KernelType kernel, bool centerTranformedData, mat data, mat transformedDataReturnMat, vec eigvalReturnVector, mat eigvecReturnMat, size_t returnDim)
+{
+ if (strcmp(nystroemMethod, "kmeans") == 0)
+ {
+ KernelPCA<KernelType, NystroemKernelRule<KernelType, KMeansSelection<>>>(kernel, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else if (strcmp(nystroemMethod, "ordered") == 0)
+ {
+ KernelPCA<KernelType, NystroemKernelRule<KernelType, OrderedSelection>>(kernel, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else if (strcmp(nystroemMethod, "random") == 0)
+ {
+ KernelPCA<KernelType, NystroemKernelRule<KernelType, RandomSelection>>(kernel, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ cout << "wrong nystroem input" << endl;
+ }
+}
+
+// TODO:
+// input: const KernelType kernel = KernelType(),
+// const bool centerTransformedData = false
+// const arma::mat & data,
+// arma::mat & transformedData,
+// arma::vec & eigval,
+// arma::mat & eigvec,
+// const size_t newDimension
+// output:
+// description:
+void kernel_pca(char const *kernel, SP_integer centerTranformedData, SP_integer useNystoem, char const *nystroemMethod,
+ double degree, double offset, double bandwidth, double scale,
+ float *dataMatArr, SP_integer dataMatSize, SP_integer dataMatRowNum,
+ float **transformedDataMatArr, SP_integer *transformedDataMatColNum, SP_integer *transformedDataMatRowNum,
+ float **eigvalArr, SP_integer *eigvalArrSize,
+ float **eigvecMatArr, SP_integer *eigvecMatColNum, SP_integer *eigvecMatRowNum,
+ SP_integer *newDimension)
+{
+ // convert the Prolog arrays to arma::mat
+ mat data = convertArrayToMat(dataMatArr, dataMatSize, dataMatRowNum);
+
+
+ // create the ReturnMat
+ mat transformedDataReturnMat;
+
+ // create the ReturnVector
+ vec eigvalReturnVector;
+
+ // create the ReturnMat
+ mat eigvecReturnMat;
+
+ size_t returnDim = 0;
+
+
+ if (strcmp(kernel, "linear") == 0)
+ {
+ LinearKernel lk;
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<LinearKernel>(nystroemMethod, lk, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<LinearKernel, NaiveKernelRule<LinearKernel>>(lk, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else if (strcmp(kernel, "polynomial") == 0)
+ {
+ PolynomialKernel pk(degree, offset);
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<PolynomialKernel>(nystroemMethod, pk, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<PolynomialKernel, NaiveKernelRule<PolynomialKernel>>(pk, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else if (strcmp(kernel, "cosine") == 0)
+ {
+ CosineDistance cd;
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<CosineDistance>(nystroemMethod, cd, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<CosineDistance, NaiveKernelRule<CosineDistance>>(cd, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else if (strcmp(kernel, "gaussian") == 0)
+ {
+ GaussianKernel gk(bandwidth);
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<GaussianKernel>(nystroemMethod, gk, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<GaussianKernel, NaiveKernelRule<GaussianKernel>>(gk, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else if (strcmp(kernel, "epanechnikov") == 0)
+ {
+ EpanechnikovKernel ek(bandwidth);
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<EpanechnikovKernel>(nystroemMethod, ek, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<EpanechnikovKernel, NaiveKernelRule<EpanechnikovKernel>>(ek, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else if (strcmp(kernel, "laplacian") == 0)
+ {
+ LaplacianKernel tk(bandwidth);
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<LaplacianKernel>(nystroemMethod, tk, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<LaplacianKernel, NaiveKernelRule<LaplacianKernel>>(tk, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else if (strcmp(kernel, "hyptan") == 0)
+ {
+ HyperbolicTangentKernel htk(scale, offset);
+ if(useNystoem == 1)
+ {
+ useNystroemKernelPCA<HyperbolicTangentKernel>(nystroemMethod, htk, (centerTranformedData == 1), data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ else
+ {
+ KernelPCA<HyperbolicTangentKernel, NaiveKernelRule<HyperbolicTangentKernel>>(htk, (centerTranformedData == 1))
+ .Apply(data, transformedDataReturnMat, eigvalReturnVector, eigvecReturnMat, returnDim);
+ }
+ }
+ else
+ {
+ cout << "wrong kernel input." << endl;
+ }
+
+ // return the transformedData Matrix
+ returnMatrixInformation(transformedDataReturnMat, transformedDataMatArr, transformedDataMatColNum, transformedDataMatRowNum);
+
+ // return the eigval Vector
+ returnVectorInformation(eigvalReturnVector, eigvalArr, eigvalArrSize);
+
+ // return the eigvec Matrix
+ returnMatrixInformation(eigvecReturnMat, eigvecMatArr, eigvecMatColNum, eigvecMatRowNum);
+
+ *newDimension = returnDim;
+}
\ No newline at end of file
diff --git a/src/methods/kernel_pca/kernel_pca.pl b/src/methods/kernel_pca/kernel_pca.pl
new file mode 100644
index 0000000000000000000000000000000000000000..55e8a944c39f3156551c240c69de9407fc845a3b
--- /dev/null
+++ b/src/methods/kernel_pca/kernel_pca.pl
@@ -0,0 +1,49 @@
+:- module(kernel_pca, [kernel_pca/20]).
+
+%% requirements of library(struct)
+:- load_files(library(str_decl),
+ [when(compile_time), if(changed)]).
+
+%% needed for using the array type
+:- use_module(library(structs)).
+:- use_module('../../helper_files/helper.pl').
+
+%% type definitions for the float array
+:- foreign_type
+ float32 = float_32,
+ float_array = array(float32).
+
+%% definitions for the connected function
+
+%% TODO:
+%% --Input--
+%% string kernel "advb","agbadf",
+%% bool name => (1)true / (0)false,
+%% bool name => (1)true / (0)false,
+%% string nystroemMethod "kmeans", "ordered", "random",
+%% float32 degree needed by polynomial,
+%% float32 offset needed by polynomial, hyptan,
+%% float32 bandwidth needed by gaussian, epanechnikov, laplacian,
+%% float32 scale needed by hyptan,
+%% mat data
+%%
+%% --Output--
+%% mat transformedData,
+%% vec eigenValues,
+%% mat eigenVectores,
+%% int newDimension
+%%
+%% --Description--
+foreign(kernel_pca, c, kernel_pca(+string, +integer, +integer, +string,
+ +float32, +float32, +float32, +float32,
+ +pointer(float_array), +integer, +integer,
+ -pointer(float_array), -integer, -integer,
+ -pointer(float_array), -integer,
+ -pointer(float_array), -integer, -integer,
+ -integer)).
+
+
+%% Defines the functions that get connected from main.cpp
+foreign_resource(kernel_pca, [kernel_pca]).
+
+:- load_foreign_resource(kernel_pca).
\ No newline at end of file
diff --git a/src/methods/kernel_pca/kernel_pca_test.pl b/src/methods/kernel_pca/kernel_pca_test.pl
new file mode 100644
index 0000000000000000000000000000000000000000..928e67ebc5d0b9b079812c67e30d9c4b11df4a81
--- /dev/null
+++ b/src/methods/kernel_pca/kernel_pca_test.pl
@@ -0,0 +1,56 @@
+:- use_module(library(plunit)).
+
+:- use_module(kernel_pca).
+:- use_module('../../helper_files/helper.pl').
+
+reset_Model :-
+ initModel(1,0,50,0.0001).
+
+:- begin_tests(lists).
+
+%% alpha tests
+test(alpha_std_init) :-
+ reset_Model,
+ alpha(0).
+test(alpha_wrong_input, fail) :-
+ reset_Model,
+ alpha(1).
+test(alpha_after_train, A =:= 9223372036854775808) :-
+ reset_Model,
+ convert_list_to_float_array([5.1,3.5,1.4,4.9,3.0,1.4,4.7,3.2,1.3,4.6,3.1,1.5],3, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize),
+ alpha(A).
+
+%% train tests
+test(correct_train) :-
+ reset_Model,
+ convert_list_to_float_array([5.1,3.5,1.4,4.9,3.0,1.4,4.7,3.2,1.3,4.6,3.1,1.5],3, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize).
+test(false_train, fail) :-
+ reset_Model,
+ convert_list_to_float_array([],3, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize).
+test(false_train2, fail) :-
+ reset_Model,
+ convert_list_to_float_array([],0, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize).
+test(false_train3, fail) :-
+ reset_Model,
+ convert_list_to_float_array([1,2],0, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize).
+test(false_train3, fail) :-
+ reset_Model,
+ convert_list_to_float_array([1,2,44,3],3, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize).
+test(false_train4) :-
+ reset_Model,
+ convert_list_to_float_array([1,2,44,3],2, array(Xsize, Xrownum, X)),
+ convert_list_to_float_array([0.2,0.2,0.2,0.2], array(Ysize, Y)),
+ train(X,Xsize, Xrownum,Y, Ysize).
+:- end_tests(lists).
\ No newline at end of file