First commit

libsvm-3.36/matlab/Makefile

@@ -0,0 +1,39 @@
+# This Makefile is used under Linux
+
+MATLABDIR ?= /usr/local/matlab
+# for Mac
+# MATLABDIR ?= /opt/local/matlab
+
+CXX ?= g++
+#CXX = g++-4.1
+CFLAGS = -Wall -Wconversion -O3 -fPIC -I$(MATLABDIR)/extern/include -I..
+
+MEX = $(MATLABDIR)/bin/mex
+MEX_OPTION = CC="$(CXX)" CXX="$(CXX)" CFLAGS="$(CFLAGS)" CXXFLAGS="$(CFLAGS)"
+# comment the following line if you use MATLAB on 32-bit computer
+MEX_OPTION += -largeArrayDims
+MEX_EXT = $(shell $(MATLABDIR)/bin/mexext)
+
+all:	matlab
+
+matlab:	binary
+
+octave:
+	@echo "please type make under Octave"
+
+binary: svmpredict.$(MEX_EXT) svmtrain.$(MEX_EXT) libsvmread.$(MEX_EXT) libsvmwrite.$(MEX_EXT)
+
+svmpredict.$(MEX_EXT):     svmpredict.c ../svm.h ../svm.cpp svm_model_matlab.c
+	$(MEX) $(MEX_OPTION) svmpredict.c ../svm.cpp svm_model_matlab.c
+
+svmtrain.$(MEX_EXT):       svmtrain.c ../svm.h ../svm.cpp svm_model_matlab.c
+	$(MEX) $(MEX_OPTION) svmtrain.c ../svm.cpp svm_model_matlab.c
+
+libsvmread.$(MEX_EXT):	libsvmread.c
+	$(MEX) $(MEX_OPTION) libsvmread.c
+
+libsvmwrite.$(MEX_EXT):	libsvmwrite.c
+	$(MEX) $(MEX_OPTION) libsvmwrite.c
+
+clean:
+	rm -f *~ *.o *.mex* *.obj

libsvm-3.36/matlab/README

@@ -0,0 +1,243 @@
+-----------------------------------------
+--- MATLAB/OCTAVE interface of LIBSVM ---
+-----------------------------------------
+
+Table of Contents
+=================
+
+- Introduction
+- Installation
+- Usage
+- Returned Model Structure
+- Other Utilities
+- Examples
+- Additional Information
+
+
+Introduction
+============
+
+This tool provides a simple interface to LIBSVM, a library for support vector
+machines (http://www.csie.ntu.edu.tw/~cjlin/libsvm). It is very easy to use as
+the usage and the way of specifying parameters are the same as that of LIBSVM.
+
+Installation
+============
+
+On Windows systems, pre-built mex files are already in the
+directory '..\windows', so please just copy them to the matlab
+directory. Now we provide binary files only for 64bit MATLAB on
+Windows. If you would like to re-build the package, please rely on the
+following steps.
+
+We recommend using make.m on both MATLAB and OCTAVE. Just type 'make'
+to build 'libsvmread.mex', 'libsvmwrite.mex', 'svmtrain.mex', and
+'svmpredict.mex'.
+
+On MATLAB or Octave:
+
+        >> make
+
+If make.m does not work on MATLAB (especially for Windows), try 'mex
+-setup' to choose a suitable compiler for mex. Make sure your compiler
+is accessible and workable. Then type 'make' to do the installation.
+
+Example:
+
+	matlab>> mex -setup
+
+MATLAB will choose the default compiler. If you have multiple compliers,
+a list is given and you can choose one from the list. For more details,
+please check the following page:
+
+https://www.mathworks.com/help/matlab/matlab_external/choose-c-or-c-compilers.html
+
+On Windows, make.m has been tested via using Visual C++.
+
+On Unix systems, if neither make.m nor 'mex -setup' works, please use
+Makefile and type 'make' in a command window. Note that we assume
+your MATLAB is installed in '/usr/local/matlab'. If not, please change
+MATLABDIR in Makefile.
+
+Example:
+        linux> make
+
+To use octave, type 'make octave':
+
+Example:
+	linux> make octave
+
+For a list of supported/compatible compilers for MATLAB, please check
+the following page:
+
+http://www.mathworks.com/support/compilers/current_release/
+
+Usage
+=====
+
+matlab> model = svmtrain(training_label_vector, training_instance_matrix [, 'libsvm_options']);
+
+        -training_label_vector:
+            An m by 1 vector of training labels (type must be double).
+        -training_instance_matrix:
+            An m by n matrix of m training instances with n features.
+            It can be dense or sparse (type must be double).
+        -libsvm_options:
+            A string of training options in the same format as that of LIBSVM.
+
+matlab> [predicted_label, accuracy, decision_values/prob_estimates] = svmpredict(testing_label_vector, testing_instance_matrix, model [, 'libsvm_options']);
+matlab> [predicted_label] = svmpredict(testing_label_vector, testing_instance_matrix, model [, 'libsvm_options']);
+
+        -testing_label_vector:
+            An m by 1 vector of prediction labels. If labels of test
+            data are unknown, simply use any random values. (type must be double)
+        -testing_instance_matrix:
+            An m by n matrix of m testing instances with n features.
+            It can be dense or sparse. (type must be double)
+        -model:
+            The output of svmtrain.
+        -libsvm_options:
+            A string of testing options in the same format as that of LIBSVM.
+
+Returned Model Structure
+========================
+
+The 'svmtrain' function returns a model which can be used for future
+prediction.  It is a structure and is organized as [Parameters, nr_class,
+totalSV, rho, Label, ProbA, ProbB, Prob_density_marks, nSV, sv_coef, SVs]:
+
+        -Parameters: parameters
+        -nr_class: number of classes; = 2 for regression/one-class svm
+        -totalSV: total #SV
+        -rho: -b of the decision function(s) wx+b
+        -Label: label of each class; empty for regression/one-class SVM
+        -sv_indices: values in [1,...,num_traning_data] to indicate SVs in the training set
+        -ProbA: pairwise probability information; empty if -b 0 or in one-class SVM
+        -ProbB: pairwise probability information; empty if -b 0 or in one-class SVM
+        -Prob_density_marks: probability information for one-class SVM; empty if -b 0 or not in one-class SVM
+        -nSV: number of SVs for each class; empty for regression/one-class SVM
+        -sv_coef: coefficients for SVs in decision functions
+        -SVs: support vectors
+
+If you do not use the option '-b 1', ProbA and ProbB are empty
+matrices. If the '-v' option is specified, cross validation is
+conducted and the returned model is just a scalar: cross-validation
+accuracy for classification and mean-squared error for regression.
+
+More details about this model can be found in LIBSVM FAQ
+(http://www.csie.ntu.edu.tw/~cjlin/libsvm/faq.html) and LIBSVM
+implementation document
+(http://www.csie.ntu.edu.tw/~cjlin/papers/libsvm.pdf).
+
+Result of Prediction
+====================
+
+The function 'svmpredict' has three outputs. The first one,
+predictd_label, is a vector of predicted labels. The second output,
+accuracy, is a vector including accuracy (for classification), mean
+squared error, and squared correlation coefficient (for regression).
+The third is a matrix containing decision values or probability
+estimates (if '-b 1' is specified). If k is the number of classes
+in training data, for decision values, each row includes results of
+predicting k(k-1)/2 binary-class SVMs. For classification, k = 1 is a
+special case. Decision value +1 is returned for each testing instance,
+instead of an empty vector. For probabilities, each row contains k values
+indicating the probability that the testing instance is in each class.
+Note that the order of classes here is the same as 'Label' field
+in the model structure.
+For one-class SVM, each row contains two elements for probabilities
+of normal instance/outlier.
+
+Other Utilities
+===============
+
+A matlab function libsvmread reads files in LIBSVM format:
+
+[label_vector, instance_matrix] = libsvmread('data.txt');
+
+Two outputs are labels and instances, which can then be used as inputs
+of svmtrain or svmpredict.
+
+A matlab function libsvmwrite writes Matlab matrix to a file in LIBSVM format:
+
+libsvmwrite('data.txt', label_vector, instance_matrix)
+
+The instance_matrix must be a sparse matrix. (type must be double)
+For 32bit and 64bit MATLAB on Windows, pre-built binary files are ready
+in the directory `..\windows', but in future releases, we will only
+include 64bit MATLAB binary files.
+
+These codes are prepared by Rong-En Fan and Kai-Wei Chang from National
+Taiwan University.
+
+Examples
+========
+
+Train and test on the provided data heart_scale:
+
+matlab> [heart_scale_label, heart_scale_inst] = libsvmread('../heart_scale');
+matlab> model = svmtrain(heart_scale_label, heart_scale_inst, '-c 1 -g 0.07');
+matlab> [predict_label, accuracy, dec_values] = svmpredict(heart_scale_label, heart_scale_inst, model); % test the training data
+
+For probability estimates, you need '-b 1' for training and testing:
+
+matlab> [heart_scale_label, heart_scale_inst] = libsvmread('../heart_scale');
+matlab> model = svmtrain(heart_scale_label, heart_scale_inst, '-c 1 -g 0.07 -b 1');
+matlab> [heart_scale_label, heart_scale_inst] = libsvmread('../heart_scale');
+matlab> [predict_label, accuracy, prob_estimates] = svmpredict(heart_scale_label, heart_scale_inst, model, '-b 1');
+
+To use precomputed kernel, you must include sample serial number as
+the first column of the training and testing data (assume your kernel
+matrix is K, # of instances is n):
+
+matlab> K1 = [(1:n)', K]; % include sample serial number as first column
+matlab> model = svmtrain(label_vector, K1, '-t 4');
+matlab> [predict_label, accuracy, dec_values] = svmpredict(label_vector, K1, model); % test the training data
+
+We give the following detailed example by splitting heart_scale into
+150 training and 120 testing data.  Constructing a linear kernel
+matrix and then using the precomputed kernel gives exactly the same
+testing error as using the LIBSVM built-in linear kernel.
+
+matlab> [heart_scale_label, heart_scale_inst] = libsvmread('../heart_scale');
+matlab>
+matlab> % Split Data
+matlab> train_data = heart_scale_inst(1:150,:);
+matlab> train_label = heart_scale_label(1:150,:);
+matlab> test_data = heart_scale_inst(151:270,:);
+matlab> test_label = heart_scale_label(151:270,:);
+matlab>
+matlab> % Linear Kernel
+matlab> model_linear = svmtrain(train_label, train_data, '-t 0');
+matlab> [predict_label_L, accuracy_L, dec_values_L] = svmpredict(test_label, test_data, model_linear);
+matlab>
+matlab> % Precomputed Kernel
+matlab> model_precomputed = svmtrain(train_label, [(1:150)', train_data*train_data'], '-t 4');
+matlab> [predict_label_P, accuracy_P, dec_values_P] = svmpredict(test_label, [(1:120)', test_data*train_data'], model_precomputed);
+matlab>
+matlab> accuracy_L % Display the accuracy using linear kernel
+matlab> accuracy_P % Display the accuracy using precomputed kernel
+
+Note that for testing, you can put anything in the
+testing_label_vector.  For more details of precomputed kernels, please
+read the section ``Precomputed Kernels'' in the README of the LIBSVM
+package.
+
+Additional Information
+======================
+
+This interface was initially written by Jun-Cheng Chen, Kuan-Jen Peng,
+Chih-Yuan Yang and Chih-Huai Cheng from Department of Computer
+Science, National Taiwan University. The current version was prepared
+by Rong-En Fan and Ting-Fan Wu. If you find this tool useful, please
+cite LIBSVM as follows
+
+Chih-Chung Chang and Chih-Jen Lin, LIBSVM : a library for support
+vector machines. ACM Transactions on Intelligent Systems and
+Technology, 2:27:1--27:27, 2011. Software available at
+http://www.csie.ntu.edu.tw/~cjlin/libsvm
+
+For any question, please contact Chih-Jen Lin <cjlin@csie.ntu.edu.tw>,
+or check the FAQ page:
+
+http://www.csie.ntu.edu.tw/~cjlin/libsvm/faq.html#/Q10:_MATLAB_interface

libsvm-3.36/matlab/libsvmread.c

@@ -0,0 +1,212 @@
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <ctype.h>
+#include <errno.h>
+
+#include "mex.h"
+
+#ifdef MX_API_VER
+#if MX_API_VER < 0x07030000
+typedef int mwIndex;
+#endif
+#endif
+#ifndef max
+#define max(x,y) (((x)>(y))?(x):(y))
+#endif
+#ifndef min
+#define min(x,y) (((x)<(y))?(x):(y))
+#endif
+
+void exit_with_help()
+{
+	mexPrintf(
+	"Usage: [label_vector, instance_matrix] = libsvmread('filename');\n"
+	);
+}
+
+static void fake_answer(int nlhs, mxArray *plhs[])
+{
+	int i;
+	for(i=0;i<nlhs;i++)
+		plhs[i] = mxCreateDoubleMatrix(0, 0, mxREAL);
+}
+
+static char *line;
+static int max_line_len;
+
+static char* readline(FILE *input)
+{
+	int len;
+
+	if(fgets(line,max_line_len,input) == NULL)
+		return NULL;
+
+	while(strrchr(line,'\n') == NULL)
+	{
+		max_line_len *= 2;
+		line = (char *) realloc(line, max_line_len);
+		len = (int) strlen(line);
+		if(fgets(line+len,max_line_len-len,input) == NULL)
+			break;
+	}
+	return line;
+}
+
+// read in a problem (in libsvm format)
+void read_problem(const char *filename, int nlhs, mxArray *plhs[])
+{
+	int max_index, min_index, inst_max_index;
+	size_t elements, k, i, l=0;
+	FILE *fp = fopen(filename,"r");
+	char *endptr;
+	mwIndex *ir, *jc;
+	double *labels, *samples;
+
+	if(fp == NULL)
+	{
+		mexPrintf("can't open input file %s\n",filename);
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	max_line_len = 1024;
+	line = (char *) malloc(max_line_len*sizeof(char));
+
+	max_index = 0;
+	min_index = 1; // our index starts from 1
+	elements = 0;
+	while(readline(fp) != NULL)
+	{
+		char *idx, *val;
+		// features
+		int index = 0;
+
+		inst_max_index = -1; // strtol gives 0 if wrong format, and precomputed kernel has <index> start from 0
+		strtok(line," \t"); // label
+		while (1)
+		{
+			idx = strtok(NULL,":"); // index:value
+			val = strtok(NULL," \t");
+			if(val == NULL)
+				break;
+
+			errno = 0;
+			index = (int) strtol(idx,&endptr,10);
+			if(endptr == idx || errno != 0 || *endptr != '\0' || index <= inst_max_index)
+			{
+				mexPrintf("Wrong input format at line %d\n",l+1);
+				fake_answer(nlhs, plhs);
+				return;
+			}
+			else
+				inst_max_index = index;
+
+			min_index = min(min_index, index);
+			elements++;
+		}
+		max_index = max(max_index, inst_max_index);
+		l++;
+	}
+	rewind(fp);
+
+	// y
+	plhs[0] = mxCreateDoubleMatrix(l, 1, mxREAL);
+	// x^T
+	if (min_index <= 0)
+		plhs[1] = mxCreateSparse(max_index-min_index+1, l, elements, mxREAL);
+	else
+		plhs[1] = mxCreateSparse(max_index, l, elements, mxREAL);
+
+	labels = mxGetPr(plhs[0]);
+	samples = mxGetPr(plhs[1]);
+	ir = mxGetIr(plhs[1]);
+	jc = mxGetJc(plhs[1]);
+
+	k=0;
+	for(i=0;i<l;i++)
+	{
+		char *idx, *val, *label;
+		jc[i] = k;
+
+		readline(fp);
+
+		label = strtok(line," \t\n");
+		if(label == NULL)
+		{
+			mexPrintf("Empty line at line %d\n",i+1);
+			fake_answer(nlhs, plhs);
+			return;
+		}
+		labels[i] = strtod(label,&endptr);
+		if(endptr == label || *endptr != '\0')
+		{
+			mexPrintf("Wrong input format at line %d\n",i+1);
+			fake_answer(nlhs, plhs);
+			return;
+		}
+
+		// features
+		while(1)
+		{
+			idx = strtok(NULL,":");
+			val = strtok(NULL," \t");
+			if(val == NULL)
+				break;
+
+			ir[k] = (mwIndex) (strtol(idx,&endptr,10) - min_index); // precomputed kernel has <index> start from 0
+
+			errno = 0;
+			samples[k] = strtod(val,&endptr);
+			if (endptr == val || errno != 0 || (*endptr != '\0' && !isspace(*endptr)))
+			{
+				mexPrintf("Wrong input format at line %d\n",i+1);
+				fake_answer(nlhs, plhs);
+				return;
+			}
+			++k;
+		}
+	}
+	jc[l] = k;
+
+	fclose(fp);
+	free(line);
+
+	{
+		mxArray *rhs[1], *lhs[1];
+		rhs[0] = plhs[1];
+		if(mexCallMATLAB(1, lhs, 1, rhs, "transpose"))
+		{
+			mexPrintf("Error: cannot transpose problem\n");
+			fake_answer(nlhs, plhs);
+			return;
+		}
+		plhs[1] = lhs[0];
+	}
+}
+
+void mexFunction( int nlhs, mxArray *plhs[],
+		int nrhs, const mxArray *prhs[] )
+{
+	#define filename_size 256
+
+	char filename[filename_size];
+
+	if(nrhs != 1 || nlhs != 2)
+	{
+		exit_with_help();
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	if(mxGetString(prhs[0], filename, filename_size) == 1){
+		mexPrintf("Error: wrong or too long filename\n");
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	read_problem(filename, nlhs, plhs);
+
+	return;
+}
+

libsvm-3.36/matlab/libsvmwrite.c

@@ -0,0 +1,119 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "mex.h"
+
+#ifdef MX_API_VER
+#if MX_API_VER < 0x07030000
+typedef int mwIndex;
+#endif
+#endif
+
+void exit_with_help()
+{
+	mexPrintf(
+	"Usage: libsvmwrite('filename', label_vector, instance_matrix);\n"
+	);
+}
+
+static void fake_answer(int nlhs, mxArray *plhs[])
+{
+	int i;
+	for(i=0;i<nlhs;i++)
+		plhs[i] = mxCreateDoubleMatrix(0, 0, mxREAL);
+}
+
+void libsvmwrite(const char *filename, const mxArray *label_vec, const mxArray *instance_mat)
+{
+	FILE *fp = fopen(filename,"w");
+	mwIndex *ir, *jc, k, low, high;
+	size_t i, l, label_vector_row_num;
+	double *samples, *labels;
+	mxArray *instance_mat_col; // instance sparse matrix in column format
+
+	if(fp ==NULL)
+	{
+		mexPrintf("can't open output file %s\n",filename);
+		return;
+	}
+
+	// transpose instance matrix
+	{
+		mxArray *prhs[1], *plhs[1];
+		prhs[0] = mxDuplicateArray(instance_mat);
+		if(mexCallMATLAB(1, plhs, 1, prhs, "transpose"))
+		{
+			mexPrintf("Error: cannot transpose instance matrix\n");
+			return;
+		}
+		instance_mat_col = plhs[0];
+		mxDestroyArray(prhs[0]);
+	}
+
+	// the number of instance
+	l = mxGetN(instance_mat_col);
+	label_vector_row_num = mxGetM(label_vec);
+
+	if(label_vector_row_num!=l)
+	{
+		mexPrintf("Length of label vector does not match # of instances.\n");
+		return;
+	}
+
+	// each column is one instance
+	labels = mxGetPr(label_vec);
+	samples = mxGetPr(instance_mat_col);
+	ir = mxGetIr(instance_mat_col);
+	jc = mxGetJc(instance_mat_col);
+
+	for(i=0;i<l;i++)
+	{
+		fprintf(fp,"%.17g", labels[i]);
+
+		low = jc[i], high = jc[i+1];
+		for(k=low;k<high;k++)
+			fprintf(fp," %lu:%g", (size_t)ir[k]+1, samples[k]);
+
+		fprintf(fp,"\n");
+	}
+
+	fclose(fp);
+	return;
+}
+
+void mexFunction( int nlhs, mxArray *plhs[],
+		int nrhs, const mxArray *prhs[] )
+{
+	if(nlhs > 0)
+	{
+		exit_with_help();
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	// Transform the input Matrix to libsvm format
+	if(nrhs == 3)
+	{
+		char filename[256];
+		if(!mxIsDouble(prhs[1]) || !mxIsDouble(prhs[2]))
+		{
+			mexPrintf("Error: label vector and instance matrix must be double\n");
+			return;
+		}
+
+		mxGetString(prhs[0], filename, mxGetN(prhs[0])+1);
+
+		if(mxIsSparse(prhs[2]))
+			libsvmwrite(filename, prhs[1], prhs[2]);
+		else
+		{
+			mexPrintf("Instance_matrix must be sparse\n");
+			return;
+		}
+	}
+	else
+	{
+		exit_with_help();
+		return;
+	}
+}

libsvm-3.36/matlab/make.m

@@ -0,0 +1,22 @@
+% This make.m is for MATLAB and OCTAVE under Windows, Mac, and Unix
+function make()
+try
+	% This part is for OCTAVE
+	if (exist ('OCTAVE_VERSION', 'builtin'))
+		mex libsvmread.c
+		mex libsvmwrite.c
+		mex -I.. svmtrain.c ../svm.cpp svm_model_matlab.c
+		mex -I.. svmpredict.c ../svm.cpp svm_model_matlab.c
+	% This part is for MATLAB
+	% Add -largeArrayDims on 64-bit machines of MATLAB
+	else
+		mex -largeArrayDims libsvmread.c
+		mex -largeArrayDims libsvmwrite.c
+		mex -I.. -largeArrayDims svmtrain.c ../svm.cpp svm_model_matlab.c
+		mex -I.. -largeArrayDims svmpredict.c ../svm.cpp svm_model_matlab.c
+	end
+catch err
+	fprintf('Error: %s failed (line %d)\n', err.stack(1).file, err.stack(1).line);
+	disp(err.message);
+	fprintf('=> Please check README for detailed instructions.\n');
+end

libsvm-3.36/matlab/svm_model_matlab.c

@@ -0,0 +1,400 @@
+#include <stdlib.h>
+#include <string.h>
+#include "svm.h"
+
+#include "mex.h"
+
+#ifdef MX_API_VER
+#if MX_API_VER < 0x07030000
+typedef int mwIndex;
+#endif
+#endif
+
+#define NUM_OF_RETURN_FIELD 12
+
+#define Malloc(type,n) (type *)malloc((n)*sizeof(type))
+
+static const char *field_names[] = {
+	"Parameters",
+	"nr_class",
+	"totalSV",
+	"rho",
+	"Label",
+	"sv_indices",
+	"ProbA",
+	"ProbB",
+	"Prob_density_marks",
+	"nSV",
+	"sv_coef",
+	"SVs"
+};
+
+const char *model_to_matlab_structure(mxArray *plhs[], int num_of_feature, struct svm_model *model)
+{
+	int i, j, n;
+	double *ptr;
+	mxArray *return_model, **rhs;
+	int out_id = 0;
+
+	rhs = (mxArray **)mxMalloc(sizeof(mxArray *)*NUM_OF_RETURN_FIELD);
+
+	// Parameters
+	rhs[out_id] = mxCreateDoubleMatrix(5, 1, mxREAL);
+	ptr = mxGetPr(rhs[out_id]);
+	ptr[0] = model->param.svm_type;
+	ptr[1] = model->param.kernel_type;
+	ptr[2] = model->param.degree;
+	ptr[3] = model->param.gamma;
+	ptr[4] = model->param.coef0;
+	out_id++;
+
+	// nr_class
+	rhs[out_id] = mxCreateDoubleMatrix(1, 1, mxREAL);
+	ptr = mxGetPr(rhs[out_id]);
+	ptr[0] = model->nr_class;
+	out_id++;
+
+	// total SV
+	rhs[out_id] = mxCreateDoubleMatrix(1, 1, mxREAL);
+	ptr = mxGetPr(rhs[out_id]);
+	ptr[0] = model->l;
+	out_id++;
+
+	// rho
+	n = model->nr_class*(model->nr_class-1)/2;
+	rhs[out_id] = mxCreateDoubleMatrix(n, 1, mxREAL);
+	ptr = mxGetPr(rhs[out_id]);
+	for(i = 0; i < n; i++)
+		ptr[i] = model->rho[i];
+	out_id++;
+
+	// Label
+	if(model->label)
+	{
+		rhs[out_id] = mxCreateDoubleMatrix(model->nr_class, 1, mxREAL);
+		ptr = mxGetPr(rhs[out_id]);
+		for(i = 0; i < model->nr_class; i++)
+			ptr[i] = model->label[i];
+	}
+	else
+		rhs[out_id] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	out_id++;
+
+	// sv_indices
+	if(model->sv_indices)
+	{
+		rhs[out_id] = mxCreateDoubleMatrix(model->l, 1, mxREAL);
+		ptr = mxGetPr(rhs[out_id]);
+		for(i = 0; i < model->l; i++)
+			ptr[i] = model->sv_indices[i];
+	}
+	else
+		rhs[out_id] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	out_id++;
+
+	// probA
+	if(model->probA != NULL)
+	{
+		rhs[out_id] = mxCreateDoubleMatrix(n, 1, mxREAL);
+		ptr = mxGetPr(rhs[out_id]);
+		for(i = 0; i < n; i++)
+			ptr[i] = model->probA[i];
+	}
+	else
+		rhs[out_id] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	out_id ++;
+
+	// probB
+	if(model->probB != NULL)
+	{
+		rhs[out_id] = mxCreateDoubleMatrix(n, 1, mxREAL);
+		ptr = mxGetPr(rhs[out_id]);
+		for(i = 0; i < n; i++)
+			ptr[i] = model->probB[i];
+	}
+	else
+		rhs[out_id] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	out_id++;
+
+	// prob_density_marks
+	if(model->prob_density_marks != NULL)
+	{
+		int nr_marks = 10;
+		rhs[out_id] = mxCreateDoubleMatrix(nr_marks, 1, mxREAL);
+		ptr = mxGetPr(rhs[out_id]);
+		for(i = 0; i < nr_marks; i++)
+			ptr[i] = model->prob_density_marks[i];
+	}
+	else
+		rhs[out_id] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	out_id++;
+
+	// nSV
+	if(model->nSV)
+	{
+		rhs[out_id] = mxCreateDoubleMatrix(model->nr_class, 1, mxREAL);
+		ptr = mxGetPr(rhs[out_id]);
+		for(i = 0; i < model->nr_class; i++)
+			ptr[i] = model->nSV[i];
+	}
+	else
+		rhs[out_id] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	out_id++;
+
+	// sv_coef
+	rhs[out_id] = mxCreateDoubleMatrix(model->l, model->nr_class-1, mxREAL);
+	ptr = mxGetPr(rhs[out_id]);
+	for(i = 0; i < model->nr_class-1; i++)
+		for(j = 0; j < model->l; j++)
+			ptr[(i*(model->l))+j] = model->sv_coef[i][j];
+	out_id++;
+
+	// SVs
+	{
+		int ir_index, nonzero_element;
+		mwIndex *ir, *jc;
+		mxArray *pprhs[1], *pplhs[1];
+
+		if(model->param.kernel_type == PRECOMPUTED)
+		{
+			nonzero_element = model->l;
+			num_of_feature = 1;
+		}
+		else
+		{
+			nonzero_element = 0;
+			for(i = 0; i < model->l; i++) {
+				j = 0;
+				while(model->SV[i][j].index != -1)
+				{
+					nonzero_element++;
+					j++;
+				}
+			}
+		}
+
+		// SV in column, easier accessing
+		rhs[out_id] = mxCreateSparse(num_of_feature, model->l, nonzero_element, mxREAL);
+		ir = mxGetIr(rhs[out_id]);
+		jc = mxGetJc(rhs[out_id]);
+		ptr = mxGetPr(rhs[out_id]);
+		jc[0] = ir_index = 0;
+		for(i = 0;i < model->l; i++)
+		{
+			if(model->param.kernel_type == PRECOMPUTED)
+			{
+				// make a (1 x model->l) matrix
+				ir[ir_index] = 0;
+				ptr[ir_index] = model->SV[i][0].value;
+				ir_index++;
+				jc[i+1] = jc[i] + 1;
+			}
+			else
+			{
+				int x_index = 0;
+				while (model->SV[i][x_index].index != -1)
+				{
+					ir[ir_index] = model->SV[i][x_index].index - 1;
+					ptr[ir_index] = model->SV[i][x_index].value;
+					ir_index++, x_index++;
+				}
+				jc[i+1] = jc[i] + x_index;
+			}
+		}
+		// transpose back to SV in row
+		pprhs[0] = rhs[out_id];
+		if(mexCallMATLAB(1, pplhs, 1, pprhs, "transpose"))
+			return "cannot transpose SV matrix";
+		rhs[out_id] = pplhs[0];
+		out_id++;
+	}
+
+	/* Create a struct matrix contains NUM_OF_RETURN_FIELD fields */
+	return_model = mxCreateStructMatrix(1, 1, NUM_OF_RETURN_FIELD, field_names);
+
+	/* Fill struct matrix with input arguments */
+	for(i = 0; i < NUM_OF_RETURN_FIELD; i++)
+		mxSetField(return_model,0,field_names[i],mxDuplicateArray(rhs[i]));
+	/* return */
+	plhs[0] = return_model;
+	mxFree(rhs);
+
+	return NULL;
+}
+
+struct svm_model *matlab_matrix_to_model(const mxArray *matlab_struct, const char **msg)
+{
+	int i, j, n, num_of_fields;
+	double *ptr;
+	int id = 0;
+	struct svm_node *x_space;
+	struct svm_model *model;
+	mxArray **rhs;
+
+	num_of_fields = mxGetNumberOfFields(matlab_struct);
+	if(num_of_fields != NUM_OF_RETURN_FIELD)
+	{
+		*msg = "number of return field is not correct";
+		return NULL;
+	}
+	rhs = (mxArray **) mxMalloc(sizeof(mxArray *)*num_of_fields);
+
+	for(i=0;i<num_of_fields;i++)
+		rhs[i] = mxGetFieldByNumber(matlab_struct, 0, i);
+
+	model = Malloc(struct svm_model, 1);
+	model->rho = NULL;
+	model->probA = NULL;
+	model->probB = NULL;
+	model->prob_density_marks = NULL;
+	model->label = NULL;
+	model->sv_indices = NULL;
+	model->nSV = NULL;
+	model->free_sv = 1; // XXX
+
+	ptr = mxGetPr(rhs[id]);
+	model->param.svm_type = (int)ptr[0];
+	model->param.kernel_type  = (int)ptr[1];
+	model->param.degree	  = (int)ptr[2];
+	model->param.gamma	  = ptr[3];
+	model->param.coef0	  = ptr[4];
+	id++;
+
+	ptr = mxGetPr(rhs[id]);
+	model->nr_class = (int)ptr[0];
+	id++;
+
+	ptr = mxGetPr(rhs[id]);
+	model->l = (int)ptr[0];
+	id++;
+
+	// rho
+	n = model->nr_class * (model->nr_class-1)/2;
+	model->rho = (double*) malloc(n*sizeof(double));
+	ptr = mxGetPr(rhs[id]);
+	for(i=0;i<n;i++)
+		model->rho[i] = ptr[i];
+	id++;
+
+	// label
+	if(mxIsEmpty(rhs[id]) == 0)
+	{
+		model->label = (int*) malloc(model->nr_class*sizeof(int));
+		ptr = mxGetPr(rhs[id]);
+		for(i=0;i<model->nr_class;i++)
+			model->label[i] = (int)ptr[i];
+	}
+	id++;
+
+	// sv_indices
+	if(mxIsEmpty(rhs[id]) == 0)
+	{
+		model->sv_indices = (int*) malloc(model->l*sizeof(int));
+		ptr = mxGetPr(rhs[id]);
+		for(i=0;i<model->l;i++)
+			model->sv_indices[i] = (int)ptr[i];
+	}
+	id++;
+
+	// probA
+	if(mxIsEmpty(rhs[id]) == 0)
+	{
+		model->probA = (double*) malloc(n*sizeof(double));
+		ptr = mxGetPr(rhs[id]);
+		for(i=0;i<n;i++)
+			model->probA[i] = ptr[i];
+	}
+	id++;
+
+	// probB
+	if(mxIsEmpty(rhs[id]) == 0)
+	{
+		model->probB = (double*) malloc(n*sizeof(double));
+		ptr = mxGetPr(rhs[id]);
+		for(i=0;i<n;i++)
+			model->probB[i] = ptr[i];
+	}
+	id++;
+
+	// prob_density_marks
+	if(mxIsEmpty(rhs[id]) == 0)
+	{
+		int nr_marks = 10;
+		model->prob_density_marks = (double*) malloc(nr_marks*sizeof(double));
+		ptr = mxGetPr(rhs[id]);
+		for(i=0;i<nr_marks;i++)
+			model->prob_density_marks[i] = ptr[i];
+	}
+	id++;
+
+	// nSV
+	if(mxIsEmpty(rhs[id]) == 0)
+	{
+		model->nSV = (int*) malloc(model->nr_class*sizeof(int));
+		ptr = mxGetPr(rhs[id]);
+		for(i=0;i<model->nr_class;i++)
+			model->nSV[i] = (int)ptr[i];
+	}
+	id++;
+
+	// sv_coef
+	ptr = mxGetPr(rhs[id]);
+	model->sv_coef = (double**) malloc((model->nr_class-1)*sizeof(double));
+	for( i=0 ; i< model->nr_class -1 ; i++ )
+		model->sv_coef[i] = (double*) malloc((model->l)*sizeof(double));
+	for(i = 0; i < model->nr_class - 1; i++)
+		for(j = 0; j < model->l; j++)
+			model->sv_coef[i][j] = ptr[i*(model->l)+j];
+	id++;
+
+	// SV
+	{
+		int sr, elements;
+		int num_samples;
+		mwIndex *ir, *jc;
+		mxArray *pprhs[1], *pplhs[1];
+
+		// transpose SV
+		pprhs[0] = rhs[id];
+		if(mexCallMATLAB(1, pplhs, 1, pprhs, "transpose"))
+		{
+			svm_free_and_destroy_model(&model);
+			*msg = "cannot transpose SV matrix";
+			return NULL;
+		}
+		rhs[id] = pplhs[0];
+
+		sr = (int)mxGetN(rhs[id]);
+
+		ptr = mxGetPr(rhs[id]);
+		ir = mxGetIr(rhs[id]);
+		jc = mxGetJc(rhs[id]);
+
+		num_samples = (int)mxGetNzmax(rhs[id]);
+
+		elements = num_samples + sr;
+
+		model->SV = (struct svm_node **) malloc(sr * sizeof(struct svm_node *));
+		x_space = (struct svm_node *)malloc(elements * sizeof(struct svm_node));
+
+		// SV is in column
+		for(i=0;i<sr;i++)
+		{
+			int low = (int)jc[i], high = (int)jc[i+1];
+			int x_index = 0;
+			model->SV[i] = &x_space[low+i];
+			for(j=low;j<high;j++)
+			{
+				model->SV[i][x_index].index = (int)ir[j] + 1;
+				model->SV[i][x_index].value = ptr[j];
+				x_index++;
+			}
+			model->SV[i][x_index].index = -1;
+		}
+
+		id++;
+	}
+	mxFree(rhs);
+
+	return model;
+}

libsvm-3.36/matlab/svm_model_matlab.h

@@ -0,0 +1,2 @@
+const char *model_to_matlab_structure(mxArray *plhs[], int num_of_feature, struct svm_model *model);
+struct svm_model *matlab_matrix_to_model(const mxArray *matlab_struct, const char **error_message);

libsvm-3.36/matlab/svmpredict.c

@@ -0,0 +1,373 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "svm.h"
+
+#include "mex.h"
+#include "svm_model_matlab.h"
+
+#ifdef MX_API_VER
+#if MX_API_VER < 0x07030000
+typedef int mwIndex;
+#endif
+#endif
+
+#define CMD_LEN 2048
+
+int print_null(const char *s,...) {return 0;}
+int (*info)(const char *fmt,...) = &mexPrintf;
+
+void read_sparse_instance(const mxArray *prhs, int index, struct svm_node *x)
+{
+	int i, j, low, high;
+	mwIndex *ir, *jc;
+	double *samples;
+
+	ir = mxGetIr(prhs);
+	jc = mxGetJc(prhs);
+	samples = mxGetPr(prhs);
+
+	// each column is one instance
+	j = 0;
+	low = (int)jc[index], high = (int)jc[index+1];
+	for(i=low;i<high;i++)
+	{
+		x[j].index = (int)ir[i] + 1;
+		x[j].value = samples[i];
+		j++;
+	}
+	x[j].index = -1;
+}
+
+static void fake_answer(int nlhs, mxArray *plhs[])
+{
+	int i;
+	for(i=0;i<nlhs;i++)
+		plhs[i] = mxCreateDoubleMatrix(0, 0, mxREAL);
+}
+
+void predict(int nlhs, mxArray *plhs[], const mxArray *prhs[], struct svm_model *model, const int predict_probability)
+{
+	int label_vector_row_num, label_vector_col_num;
+	int feature_number, testing_instance_number;
+	int instance_index;
+	double *ptr_instance, *ptr_label, *ptr_predict_label;
+	double *ptr_prob_estimates, *ptr_dec_values, *ptr;
+	struct svm_node *x;
+	mxArray *pplhs[1]; // transposed instance sparse matrix
+	mxArray *tplhs[3]; // temporary storage for plhs[]
+
+	int correct = 0;
+	int total = 0;
+	double error = 0;
+	double sump = 0, sumt = 0, sumpp = 0, sumtt = 0, sumpt = 0;
+
+	int svm_type=svm_get_svm_type(model);
+	int nr_class=svm_get_nr_class(model);
+	double *prob_estimates=NULL;
+
+	// prhs[1] = testing instance matrix
+	feature_number = (int)mxGetN(prhs[1]);
+	testing_instance_number = (int)mxGetM(prhs[1]);
+	label_vector_row_num = (int)mxGetM(prhs[0]);
+	label_vector_col_num = (int)mxGetN(prhs[0]);
+
+	if(label_vector_row_num!=testing_instance_number)
+	{
+		mexPrintf("Length of label vector does not match # of instances.\n");
+		fake_answer(nlhs, plhs);
+		return;
+	}
+	if(label_vector_col_num!=1)
+	{
+		mexPrintf("label (1st argument) should be a vector (# of column is 1).\n");
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	ptr_instance = mxGetPr(prhs[1]);
+	ptr_label    = mxGetPr(prhs[0]);
+
+	// transpose instance matrix
+	if(mxIsSparse(prhs[1]))
+	{
+		if(model->param.kernel_type == PRECOMPUTED)
+		{
+			// precomputed kernel requires dense matrix, so we make one
+			mxArray *rhs[1], *lhs[1];
+			rhs[0] = mxDuplicateArray(prhs[1]);
+			if(mexCallMATLAB(1, lhs, 1, rhs, "full"))
+			{
+				mexPrintf("Error: cannot full testing instance matrix\n");
+				fake_answer(nlhs, plhs);
+				return;
+			}
+			ptr_instance = mxGetPr(lhs[0]);
+			mxDestroyArray(rhs[0]);
+		}
+		else
+		{
+			mxArray *pprhs[1];
+			pprhs[0] = mxDuplicateArray(prhs[1]);
+			if(mexCallMATLAB(1, pplhs, 1, pprhs, "transpose"))
+			{
+				mexPrintf("Error: cannot transpose testing instance matrix\n");
+				fake_answer(nlhs, plhs);
+				return;
+			}
+		}
+	}
+
+	if(predict_probability)
+	{
+		if(svm_type==NU_SVR || svm_type==EPSILON_SVR)
+			info("Prob. model for test data: target value = predicted value + z,\nz: Laplace distribution e^(-|z|/sigma)/(2sigma),sigma=%g\n",svm_get_svr_probability(model));
+		else
+			prob_estimates = (double *) malloc(nr_class*sizeof(double));
+	}
+
+	tplhs[0] = mxCreateDoubleMatrix(testing_instance_number, 1, mxREAL);
+	if(predict_probability)
+	{
+		// prob estimates are in plhs[2]
+		if(svm_type==C_SVC || svm_type==NU_SVC || svm_type==ONE_CLASS)
+		{
+			// nr_class = 2 for ONE_CLASS
+			tplhs[2] = mxCreateDoubleMatrix(testing_instance_number, nr_class, mxREAL);
+		}
+		else
+			tplhs[2] = mxCreateDoubleMatrix(0, 0, mxREAL);
+	}
+	else
+	{
+		// decision values are in plhs[2]
+		if(svm_type == ONE_CLASS ||
+		   svm_type == EPSILON_SVR ||
+		   svm_type == NU_SVR ||
+		   nr_class == 1) // if only one class in training data, decision values are still returned.
+			tplhs[2] = mxCreateDoubleMatrix(testing_instance_number, 1, mxREAL);
+		else
+			tplhs[2] = mxCreateDoubleMatrix(testing_instance_number, nr_class*(nr_class-1)/2, mxREAL);
+	}
+
+	ptr_predict_label = mxGetPr(tplhs[0]);
+	ptr_prob_estimates = mxGetPr(tplhs[2]);
+	ptr_dec_values = mxGetPr(tplhs[2]);
+	x = (struct svm_node*)malloc((feature_number+1)*sizeof(struct svm_node) );
+	for(instance_index=0;instance_index<testing_instance_number;instance_index++)
+	{
+		int i;
+		double target_label, predict_label;
+
+		target_label = ptr_label[instance_index];
+
+		if(mxIsSparse(prhs[1]) && model->param.kernel_type != PRECOMPUTED) // prhs[1]^T is still sparse
+			read_sparse_instance(pplhs[0], instance_index, x);
+		else
+		{
+			for(i=0;i<feature_number;i++)
+			{
+				x[i].index = i+1;
+				x[i].value = ptr_instance[testing_instance_number*i+instance_index];
+			}
+			x[feature_number].index = -1;
+		}
+
+		if(predict_probability)
+		{
+			if(svm_type==C_SVC || svm_type==NU_SVC || svm_type==ONE_CLASS)
+			{
+				predict_label = svm_predict_probability(model, x, prob_estimates);
+				ptr_predict_label[instance_index] = predict_label;
+				for(i=0;i<nr_class;i++)
+					ptr_prob_estimates[instance_index + i * testing_instance_number] = prob_estimates[i];
+			} else {
+				predict_label = svm_predict(model,x);
+				ptr_predict_label[instance_index] = predict_label;
+			}
+		}
+		else
+		{
+			if(svm_type == ONE_CLASS ||
+			   svm_type == EPSILON_SVR ||
+			   svm_type == NU_SVR)
+			{
+				double res;
+				predict_label = svm_predict_values(model, x, &res);
+				ptr_dec_values[instance_index] = res;
+			}
+			else
+			{
+				double *dec_values = (double *) malloc(sizeof(double) * nr_class*(nr_class-1)/2);
+				predict_label = svm_predict_values(model, x, dec_values);
+				if(nr_class == 1)
+					ptr_dec_values[instance_index] = 1;
+				else
+					for(i=0;i<(nr_class*(nr_class-1))/2;i++)
+						ptr_dec_values[instance_index + i * testing_instance_number] = dec_values[i];
+				free(dec_values);
+			}
+			ptr_predict_label[instance_index] = predict_label;
+		}
+
+		if(predict_label == target_label)
+			++correct;
+		error += (predict_label-target_label)*(predict_label-target_label);
+		sump += predict_label;
+		sumt += target_label;
+		sumpp += predict_label*predict_label;
+		sumtt += target_label*target_label;
+		sumpt += predict_label*target_label;
+		++total;
+	}
+	if(svm_type==NU_SVR || svm_type==EPSILON_SVR)
+	{
+		info("Mean squared error = %g (regression)\n",error/total);
+		info("Squared correlation coefficient = %g (regression)\n",
+			((total*sumpt-sump*sumt)*(total*sumpt-sump*sumt))/
+			((total*sumpp-sump*sump)*(total*sumtt-sumt*sumt))
+			);
+	}
+	else
+		info("Accuracy = %g%% (%d/%d) (classification)\n",
+			(double)correct/total*100,correct,total);
+
+	// return accuracy, mean squared error, squared correlation coefficient
+	tplhs[1] = mxCreateDoubleMatrix(3, 1, mxREAL);
+	ptr = mxGetPr(tplhs[1]);
+	ptr[0] = (double)correct/total*100;
+	ptr[1] = error/total;
+	ptr[2] = ((total*sumpt-sump*sumt)*(total*sumpt-sump*sumt))/
+				((total*sumpp-sump*sump)*(total*sumtt-sumt*sumt));
+
+	free(x);
+	if(prob_estimates != NULL)
+		free(prob_estimates);
+
+	switch(nlhs)
+	{
+		case 3:
+			plhs[2] = tplhs[2];
+			plhs[1] = tplhs[1];
+		case 1:
+		case 0:
+			plhs[0] = tplhs[0];
+	}
+}
+
+void exit_with_help()
+{
+	mexPrintf(
+		"Usage: [predicted_label, accuracy, decision_values/prob_estimates] = svmpredict(testing_label_vector, testing_instance_matrix, model, 'libsvm_options')\n"
+		"       [predicted_label] = svmpredict(testing_label_vector, testing_instance_matrix, model, 'libsvm_options')\n"
+		"Parameters:\n"
+		"  model: SVM model structure from svmtrain.\n"
+		"  libsvm_options:\n"
+		"    -b probability_estimates: whether to predict probability estimates, 0 or 1 (default 0); one-class SVM not supported yet\n"
+		"    -q : quiet mode (no outputs)\n"
+		"Returns:\n"
+		"  predicted_label: SVM prediction output vector.\n"
+		"  accuracy: a vector with accuracy, mean squared error, squared correlation coefficient.\n"
+		"  prob_estimates: If selected, probability estimate vector.\n"
+	);
+}
+
+void mexFunction( int nlhs, mxArray *plhs[],
+		 int nrhs, const mxArray *prhs[] )
+{
+	int prob_estimate_flag = 0;
+	struct svm_model *model;
+	info = &mexPrintf;
+
+	if(nlhs == 2 || nlhs > 3 || nrhs > 4 || nrhs < 3)
+	{
+		exit_with_help();
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	if(!mxIsDouble(prhs[0]) || !mxIsDouble(prhs[1])) {
+		mexPrintf("Error: label vector and instance matrix must be double\n");
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	if(mxIsStruct(prhs[2]))
+	{
+		const char *error_msg;
+
+		// parse options
+		if(nrhs==4)
+		{
+			int i, argc = 1;
+			char cmd[CMD_LEN], *argv[CMD_LEN/2];
+
+			// put options in argv[]
+			mxGetString(prhs[3], cmd,  mxGetN(prhs[3]) + 1);
+			if((argv[argc] = strtok(cmd, " ")) != NULL)
+				while((argv[++argc] = strtok(NULL, " ")) != NULL)
+					;
+
+			for(i=1;i<argc;i++)
+			{
+				if(argv[i][0] != '-') break;
+				if((++i>=argc) && argv[i-1][1] != 'q')
+				{
+					exit_with_help();
+					fake_answer(nlhs, plhs);
+					return;
+				}
+				switch(argv[i-1][1])
+				{
+					case 'b':
+						prob_estimate_flag = atoi(argv[i]);
+						break;
+					case 'q':
+						i--;
+						info = &print_null;
+						break;
+					default:
+						mexPrintf("Unknown option: -%c\n", argv[i-1][1]);
+						exit_with_help();
+						fake_answer(nlhs, plhs);
+						return;
+				}
+			}
+		}
+
+		model = matlab_matrix_to_model(prhs[2], &error_msg);
+		if (model == NULL)
+		{
+			mexPrintf("Error: can't read model: %s\n", error_msg);
+			fake_answer(nlhs, plhs);
+			return;
+		}
+
+		if(prob_estimate_flag)
+		{
+			if(svm_check_probability_model(model)==0)
+			{
+				mexPrintf("Model does not support probabiliy estimates\n");
+				fake_answer(nlhs, plhs);
+				svm_free_and_destroy_model(&model);
+				return;
+			}
+		}
+		else
+		{
+			if(svm_check_probability_model(model)!=0)
+				info("Model supports probability estimates, but disabled in predicton.\n");
+		}
+
+		predict(nlhs, plhs, prhs, model, prob_estimate_flag);
+		// destroy model
+		svm_free_and_destroy_model(&model);
+	}
+	else
+	{
+		mexPrintf("model file should be a struct array\n");
+		fake_answer(nlhs, plhs);
+	}
+
+	return;
+}

libsvm-3.36/matlab/svmtrain.c

@@ -0,0 +1,495 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ctype.h>
+#include "svm.h"
+
+#include "mex.h"
+#include "svm_model_matlab.h"
+
+#ifdef MX_API_VER
+#if MX_API_VER < 0x07030000
+typedef int mwIndex;
+#endif
+#endif
+
+#define CMD_LEN 2048
+#define Malloc(type,n) (type *)malloc((n)*sizeof(type))
+
+void print_null(const char *s) {}
+void print_string_matlab(const char *s) {mexPrintf(s);}
+
+void exit_with_help()
+{
+	mexPrintf(
+	"Usage: model = svmtrain(training_label_vector, training_instance_matrix, 'libsvm_options');\n"
+	"libsvm_options:\n"
+	"-s svm_type : set type of SVM (default 0)\n"
+	"	0 -- C-SVC		(multi-class classification)\n"
+	"	1 -- nu-SVC		(multi-class classification)\n"
+	"	2 -- one-class SVM\n"
+	"	3 -- epsilon-SVR	(regression)\n"
+	"	4 -- nu-SVR		(regression)\n"
+	"-t kernel_type : set type of kernel function (default 2)\n"
+	"	0 -- linear: u'*v\n"
+	"	1 -- polynomial: (gamma*u'*v + coef0)^degree\n"
+	"	2 -- radial basis function: exp(-gamma*|u-v|^2)\n"
+	"	3 -- sigmoid: tanh(gamma*u'*v + coef0)\n"
+	"	4 -- precomputed kernel (kernel values in training_instance_matrix)\n"
+	"-d degree : set degree in kernel function (default 3)\n"
+	"-g gamma : set gamma in kernel function (default 1/num_features)\n"
+	"-r coef0 : set coef0 in kernel function (default 0)\n"
+	"-c cost : set the parameter C of C-SVC, epsilon-SVR, and nu-SVR (default 1)\n"
+	"-n nu : set the parameter nu of nu-SVC, one-class SVM, and nu-SVR (default 0.5)\n"
+	"-p epsilon : set the epsilon in loss function of epsilon-SVR (default 0.1)\n"
+	"-m cachesize : set cache memory size in MB (default 100)\n"
+	"-e epsilon : set tolerance of termination criterion (default 0.001)\n"
+	"-h shrinking : whether to use the shrinking heuristics, 0 or 1 (default 1)\n"
+	"-b probability_estimates : whether to train a SVC or SVR model for probability estimates, 0 or 1 (default 0)\n"
+	"-wi weight : set the parameter C of class i to weight*C, for C-SVC (default 1)\n"
+	"-v n: n-fold cross validation mode\n"
+	"-q : quiet mode (no outputs)\n"
+	);
+}
+
+// svm arguments
+struct svm_parameter param;		// set by parse_command_line
+struct svm_problem prob;		// set by read_problem
+struct svm_model *model;
+struct svm_node *x_space;
+int cross_validation;
+int nr_fold;
+
+
+double do_cross_validation()
+{
+	int i;
+	int total_correct = 0;
+	double total_error = 0;
+	double sumv = 0, sumy = 0, sumvv = 0, sumyy = 0, sumvy = 0;
+	double *target = Malloc(double,prob.l);
+	double retval = 0.0;
+
+	svm_cross_validation(&prob,&param,nr_fold,target);
+	if(param.svm_type == EPSILON_SVR ||
+	   param.svm_type == NU_SVR)
+	{
+		for(i=0;i<prob.l;i++)
+		{
+			double y = prob.y[i];
+			double v = target[i];
+			total_error += (v-y)*(v-y);
+			sumv += v;
+			sumy += y;
+			sumvv += v*v;
+			sumyy += y*y;
+			sumvy += v*y;
+		}
+		mexPrintf("Cross Validation Mean squared error = %g\n",total_error/prob.l);
+		mexPrintf("Cross Validation Squared correlation coefficient = %g\n",
+			((prob.l*sumvy-sumv*sumy)*(prob.l*sumvy-sumv*sumy))/
+			((prob.l*sumvv-sumv*sumv)*(prob.l*sumyy-sumy*sumy))
+			);
+		retval = total_error/prob.l;
+	}
+	else
+	{
+		for(i=0;i<prob.l;i++)
+			if(target[i] == prob.y[i])
+				++total_correct;
+		mexPrintf("Cross Validation Accuracy = %g%%\n",100.0*total_correct/prob.l);
+		retval = 100.0*total_correct/prob.l;
+	}
+	free(target);
+	return retval;
+}
+
+// nrhs should be 3
+int parse_command_line(int nrhs, const mxArray *prhs[], char *model_file_name)
+{
+	int i, argc = 1;
+	char cmd[CMD_LEN];
+	char *argv[CMD_LEN/2];
+	void (*print_func)(const char *) = print_string_matlab;	// default printing to matlab display
+
+	// default values
+	param.svm_type = C_SVC;
+	param.kernel_type = RBF;
+	param.degree = 3;
+	param.gamma = 0;	// 1/num_features
+	param.coef0 = 0;
+	param.nu = 0.5;
+	param.cache_size = 100;
+	param.C = 1;
+	param.eps = 1e-3;
+	param.p = 0.1;
+	param.shrinking = 1;
+	param.probability = 0;
+	param.nr_weight = 0;
+	param.weight_label = NULL;
+	param.weight = NULL;
+	cross_validation = 0;
+
+	if(nrhs <= 1)
+		return 1;
+
+	if(nrhs > 2)
+	{
+		// put options in argv[]
+		mxGetString(prhs[2], cmd, mxGetN(prhs[2]) + 1);
+		if((argv[argc] = strtok(cmd, " ")) != NULL)
+			while((argv[++argc] = strtok(NULL, " ")) != NULL)
+				;
+	}
+
+	// parse options
+	for(i=1;i<argc;i++)
+	{
+		if(argv[i][0] != '-') break;
+		++i;
+		if(i>=argc && argv[i-1][1] != 'q')	// since option -q has no parameter
+			return 1;
+		switch(argv[i-1][1])
+		{
+			case 's':
+				param.svm_type = atoi(argv[i]);
+				break;
+			case 't':
+				param.kernel_type = atoi(argv[i]);
+				break;
+			case 'd':
+				param.degree = atoi(argv[i]);
+				break;
+			case 'g':
+				param.gamma = atof(argv[i]);
+				break;
+			case 'r':
+				param.coef0 = atof(argv[i]);
+				break;
+			case 'n':
+				param.nu = atof(argv[i]);
+				break;
+			case 'm':
+				param.cache_size = atof(argv[i]);
+				break;
+			case 'c':
+				param.C = atof(argv[i]);
+				break;
+			case 'e':
+				param.eps = atof(argv[i]);
+				break;
+			case 'p':
+				param.p = atof(argv[i]);
+				break;
+			case 'h':
+				param.shrinking = atoi(argv[i]);
+				break;
+			case 'b':
+				param.probability = atoi(argv[i]);
+				break;
+			case 'q':
+				print_func = &print_null;
+				i--;
+				break;
+			case 'v':
+				cross_validation = 1;
+				nr_fold = atoi(argv[i]);
+				if(nr_fold < 2)
+				{
+					mexPrintf("n-fold cross validation: n must >= 2\n");
+					return 1;
+				}
+				break;
+			case 'w':
+				++param.nr_weight;
+				param.weight_label = (int *)realloc(param.weight_label,sizeof(int)*param.nr_weight);
+				param.weight = (double *)realloc(param.weight,sizeof(double)*param.nr_weight);
+				param.weight_label[param.nr_weight-1] = atoi(&argv[i-1][2]);
+				param.weight[param.nr_weight-1] = atof(argv[i]);
+				break;
+			default:
+				mexPrintf("Unknown option -%c\n", argv[i-1][1]);
+				return 1;
+		}
+	}
+
+	svm_set_print_string_function(print_func);
+
+	return 0;
+}
+
+// read in a problem (in svmlight format)
+int read_problem_dense(const mxArray *label_vec, const mxArray *instance_mat)
+{
+	// using size_t due to the output type of matlab functions
+	size_t i, j, k, l;
+	size_t elements, max_index, sc, label_vector_row_num;
+	double *samples, *labels;
+
+	prob.x = NULL;
+	prob.y = NULL;
+	x_space = NULL;
+
+	labels = mxGetPr(label_vec);
+	samples = mxGetPr(instance_mat);
+	sc = mxGetN(instance_mat);
+
+	elements = 0;
+	// number of instances
+	l = mxGetM(instance_mat);
+	label_vector_row_num = mxGetM(label_vec);
+	prob.l = (int)l;
+
+	if(label_vector_row_num!=l)
+	{
+		mexPrintf("Length of label vector does not match # of instances.\n");
+		return -1;
+	}
+
+	if(param.kernel_type == PRECOMPUTED)
+		elements = l * (sc + 1);
+	else
+	{
+		for(i = 0; i < l; i++)
+		{
+			for(k = 0; k < sc; k++)
+				if(samples[k * l + i] != 0)
+					elements++;
+			// count the '-1' element
+			elements++;
+		}
+	}
+
+	prob.y = Malloc(double,l);
+	prob.x = Malloc(struct svm_node *,l);
+	x_space = Malloc(struct svm_node, elements);
+
+	max_index = sc;
+	j = 0;
+	for(i = 0; i < l; i++)
+	{
+		prob.x[i] = &x_space[j];
+		prob.y[i] = labels[i];
+
+		for(k = 0; k < sc; k++)
+		{
+			if(param.kernel_type == PRECOMPUTED || samples[k * l + i] != 0)
+			{
+				x_space[j].index = (int)k + 1;
+				x_space[j].value = samples[k * l + i];
+				j++;
+			}
+		}
+		x_space[j++].index = -1;
+	}
+
+	if(param.gamma == 0 && max_index > 0)
+		param.gamma = (double)(1.0/max_index);
+
+	if(param.kernel_type == PRECOMPUTED)
+		for(i=0;i<l;i++)
+		{
+			if((int)prob.x[i][0].value <= 0 || (int)prob.x[i][0].value > (int)max_index)
+			{
+				mexPrintf("Wrong input format: sample_serial_number out of range\n");
+				return -1;
+			}
+		}
+
+	return 0;
+}
+
+int read_problem_sparse(const mxArray *label_vec, const mxArray *instance_mat)
+{
+	mwIndex *ir, *jc, low, high, k;
+	// using size_t due to the output type of matlab functions
+	size_t i, j, l, elements, max_index, label_vector_row_num;
+	mwSize num_samples;
+	double *samples, *labels;
+	mxArray *instance_mat_col; // transposed instance sparse matrix
+
+	prob.x = NULL;
+	prob.y = NULL;
+	x_space = NULL;
+
+	// transpose instance matrix
+	{
+		mxArray *prhs[1], *plhs[1];
+		prhs[0] = mxDuplicateArray(instance_mat);
+		if(mexCallMATLAB(1, plhs, 1, prhs, "transpose"))
+		{
+			mexPrintf("Error: cannot transpose training instance matrix\n");
+			return -1;
+		}
+		instance_mat_col = plhs[0];
+		mxDestroyArray(prhs[0]);
+	}
+
+	// each column is one instance
+	labels = mxGetPr(label_vec);
+	samples = mxGetPr(instance_mat_col);
+	ir = mxGetIr(instance_mat_col);
+	jc = mxGetJc(instance_mat_col);
+
+	num_samples = mxGetNzmax(instance_mat_col);
+
+	// number of instances
+	l = mxGetN(instance_mat_col);
+	label_vector_row_num = mxGetM(label_vec);
+	prob.l = (int) l;
+
+	if(label_vector_row_num!=l)
+	{
+		mexPrintf("Length of label vector does not match # of instances.\n");
+		return -1;
+	}
+
+	elements = num_samples + l;
+	max_index = mxGetM(instance_mat_col);
+
+	prob.y = Malloc(double,l);
+	prob.x = Malloc(struct svm_node *,l);
+	x_space = Malloc(struct svm_node, elements);
+
+	j = 0;
+	for(i=0;i<l;i++)
+	{
+		prob.x[i] = &x_space[j];
+		prob.y[i] = labels[i];
+		low = jc[i], high = jc[i+1];
+		for(k=low;k<high;k++)
+		{
+			x_space[j].index = (int)ir[k] + 1;
+			x_space[j].value = samples[k];
+			j++;
+	 	}
+		x_space[j++].index = -1;
+	}
+
+	if(param.gamma == 0 && max_index > 0)
+		param.gamma = (double)(1.0/max_index);
+
+	return 0;
+}
+
+static void fake_answer(int nlhs, mxArray *plhs[])
+{
+	int i;
+	for(i=0;i<nlhs;i++)
+		plhs[i] = mxCreateDoubleMatrix(0, 0, mxREAL);
+}
+
+// Interface function of matlab
+// now assume prhs[0]: label prhs[1]: features
+void mexFunction( int nlhs, mxArray *plhs[],
+		int nrhs, const mxArray *prhs[] )
+{
+	const char *error_msg;
+
+	// fix random seed to have same results for each run
+	// (for cross validation and probability estimation)
+	srand(1);
+
+	if(nlhs > 1)
+	{
+		exit_with_help();
+		fake_answer(nlhs, plhs);
+		return;
+	}
+
+	// Transform the input Matrix to libsvm format
+	if(nrhs > 1 && nrhs < 4)
+	{
+		int err;
+
+		if(!mxIsDouble(prhs[0]) || !mxIsDouble(prhs[1]))
+		{
+			mexPrintf("Error: label vector and instance matrix must be double\n");
+			fake_answer(nlhs, plhs);
+			return;
+		}
+
+		if(mxIsSparse(prhs[0]))
+		{
+			mexPrintf("Error: label vector should not be in sparse format\n");
+			fake_answer(nlhs, plhs);
+			return;
+		}
+
+		if(parse_command_line(nrhs, prhs, NULL))
+		{
+			exit_with_help();
+			svm_destroy_param(&param);
+			fake_answer(nlhs, plhs);
+			return;
+		}
+
+		if(mxIsSparse(prhs[1]))
+		{
+			if(param.kernel_type == PRECOMPUTED)
+			{
+				// precomputed kernel requires dense matrix, so we make one
+				mxArray *rhs[1], *lhs[1];
+
+				rhs[0] = mxDuplicateArray(prhs[1]);
+				if(mexCallMATLAB(1, lhs, 1, rhs, "full"))
+				{
+					mexPrintf("Error: cannot generate a full training instance matrix\n");
+					svm_destroy_param(&param);
+					fake_answer(nlhs, plhs);
+					return;
+				}
+				err = read_problem_dense(prhs[0], lhs[0]);
+				mxDestroyArray(lhs[0]);
+				mxDestroyArray(rhs[0]);
+			}
+			else
+				err = read_problem_sparse(prhs[0], prhs[1]);
+		}
+		else
+			err = read_problem_dense(prhs[0], prhs[1]);
+
+		// svmtrain's original code
+		error_msg = svm_check_parameter(&prob, &param);
+
+		if(err || error_msg)
+		{
+			if (error_msg != NULL)
+				mexPrintf("Error: %s\n", error_msg);
+			svm_destroy_param(&param);
+			free(prob.y);
+			free(prob.x);
+			free(x_space);
+			fake_answer(nlhs, plhs);
+			return;
+		}
+
+		if(cross_validation)
+		{
+			double *ptr;
+			plhs[0] = mxCreateDoubleMatrix(1, 1, mxREAL);
+			ptr = mxGetPr(plhs[0]);
+			ptr[0] = do_cross_validation();
+		}
+		else
+		{
+			int nr_feat = (int)mxGetN(prhs[1]);
+			const char *error_msg;
+			model = svm_train(&prob, &param);
+			error_msg = model_to_matlab_structure(plhs, nr_feat, model);
+			if(error_msg)
+				mexPrintf("Error: can't convert libsvm model to matrix structure: %s\n", error_msg);
+			svm_free_and_destroy_model(&model);
+		}
+		svm_destroy_param(&param);
+		free(prob.y);
+		free(prob.x);
+		free(x_space);
+	}
+	else
+	{
+		exit_with_help();
+		fake_answer(nlhs, plhs);
+		return;
+	}
+}