svmtrain.c

Go to the documentation of this file.

#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)
{
  int i, j, k;
  int 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 = (int)mxGetN(instance_mat);

  elements = 0;
  // the number of instance
  prob.l = (int)mxGetM(instance_mat);
  label_vector_row_num = (int)mxGetM(label_vec);

  if (label_vector_row_num != prob.l)
  {
    mexPrintf("Length of label vector does not match # of instances.\n");
    return -1;
  }

  if (param.kernel_type == PRECOMPUTED)
    elements = prob.l * (sc + 1);
  else
  {
    for (i = 0; i < prob.l; i++)
    {
      for (k = 0; k < sc; k++)
        if (samples[k * prob.l + i] != 0)
          elements++;
      // count the '-1' element
      elements++;
    }
  }

  prob.y = Malloc(double, prob.l);
  prob.x = Malloc(struct svm_node *, prob.l);
  x_space = Malloc(struct svm_node, elements);

  max_index = sc;
  j = 0;
  for (i = 0; i < prob.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 * prob.l + i] != 0)
      {
        x_space[j].index = k + 1;
        x_space[j].value = samples[k * prob.l + i];
        j++;
      }
    }
    x_space[j++].index = -1;
  }

  if (param.gamma == 0 && max_index > 0)
    param.gamma = 1.0 / max_index;

  if (param.kernel_type == PRECOMPUTED)
    for (i = 0; i < prob.l; i++)
    {
      if ((int)prob.x[i][0].value <= 0 || (int)prob.x[i][0].value > 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)
{
  int i, j, k, low, high;
  mwIndex *ir, *jc;
  int elements, max_index, num_samples, label_vector_row_num;
  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 = (int)mxGetNzmax(instance_mat_col);

  // the number of instance
  prob.l = (int)mxGetN(instance_mat_col);
  label_vector_row_num = (int)mxGetM(label_vec);

  if (label_vector_row_num != prob.l)
  {
    mexPrintf("Length of label vector does not match # of instances.\n");
    return -1;
  }

  elements = num_samples + prob.l;
  max_index = (int)mxGetM(instance_mat_col);

  prob.y = Malloc(double, prob.l);
  prob.x = Malloc(struct svm_node *, prob.l);
  x_space = Malloc(struct svm_node, elements);

  j = 0;
  for (i = 0; i < prob.l; i++)
  {
    prob.x[i] = &x_space[j];
    prob.y[i] = labels[i];
    low = (int)jc[i], high = (int)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 = 1.0 / max_index;

  return 0;
}

static void fake_answer(mxArray *plhs[])
{
  plhs[0] = 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);

  // 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(plhs);
      return;
    }

    if (parse_command_line(nrhs, prhs, NULL))
    {
      exit_with_help();
      svm_destroy_param(&param);
      fake_answer(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(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(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(plhs);
    return;
  }
}