Framework.cpp

#include "Framework.h"
#include <math.h>

Framework::Framework(Network *g) {
  // Initialize the cost matrix
  int no_nodes = g->get_no_nodes();
  cost_matrix = Constants::cost_matrix;
  // for (int i = 0; i < no_nodes; i++) {
  //   vector<double> vec(no_nodes);
  //   for (int j = 0; j < no_nodes; j++) {
  //     double r = rand() % 3 + 1;
  //     vec[j] = r;
  //   }
  //   cost_matrix.push_back(vec);
  // }

  this->g = g;
  no_samples = ceil(6.75 * ((double)(g->get_no_nodes())) /
                    (Constants::EPS * Constants::EPS));
}

Framework::~Framework() {}

double Framework::estimate_test(const kseeds &seeds, uint n) {
  double re = 0.0;
  vector<bool> a(n, false);

#pragma omp parallel for
  for (int i = 0; i < n; ++i) {
    a[i] = (g->sample_influence_linear_threshold(seeds) == 1);
  }

  for (bool b : a)
    if (b)
      re += 1.0;

  return re / n * g->get_no_nodes();
}

double Framework::estimate_influence(const kseeds &seeds) {
  if (Constants::DATA == Social) {
    double re = 0.0;
    uint need_sample = 100; // Hard code need_sample
    vector<bool> a(need_sample, false);

#pragma omp parallel for
    for (int i = 0; i < need_sample; ++i) {
      a[i] = (g->sample_influence_linear_threshold(seeds) == 1);
    }

    for (bool b : a)
      if (b)
        re += 1.0;

    return re / need_sample * g->get_no_nodes();
  } else {
    return g->get_entropy(seeds);
  }
}