void FindDialog::OnOK()
{
CString str;
CEdit *pED = (CEdit*) GetDlgItem(IDC_FO_EDIT);
pED->GetWindowText(m_find_expression);
//
// Perform a synax check, and if it is bogus display a message
// box.
//
OrganismFinder ofc(m_find_expression, true);
if( ofc.error ) {
str = "Invalid Find Expression: " + ofc.error_message;
AfxMessageBox(str, MB_OK, 0);
} else {
CDialog::OnOK();
}
}
int main (int argc, char* argv[])
{
//FormMCLInputFile("mcl_input");
std::vector<std::shared_ptr<wv::Point>> points;
std::vector<std::string> answers;
log_netw->info("Hello world");
if (!ex::readDataSet("../data/compound", NumDimensionsSynthetic, points, answers))
{
std::cerr << "readIrisDataSet function works incorrect" << std::endl;
}
//0x3x14
std::string output_filename = "points";
//good result
//0x2x4
//nn::ESoinn ns(NumDimensionsSynthetic, 100 /*age_max*/, 50 /*lambda*/, 0.7 /*C1*/, 0.1 /*C2*/); //Круто для s2 работает
//nn::ESoinn ns(NumDimensionsSynthetic, 50 /*age_max*/, 25 /*lambda*/, 0.1 /*C1*/, 0.05 /*C2*/); //Compound dataset mcl 1.2
//nn::ESoinn ns(NumDimensionsSynthetic, 25 /*age_max*/, 26 /*lambda*/, 0.1 /*C1*/, 0.05 /*C2*/); //Compound dataset mcl 1.2
//nn::ESoinn ns(NumDimensionsSynthetic, 100 /*age_max*/, 50 /*lambda*/, 0.9 /*C1*/, 0.9 /*C2*/); //S2 dataset mcl 1.2 10,20 итераций
//nn::ESoinn ns(NumDimensionsSynthetic, 15 /*age_max*/, 30 /*lambda*/, 0.9 /*C1*/, 0.9 /*C2*/); //S2 dataset mcl 1.2 10,20 итераций
//----------------------------------------------------------------------------
//nn::ESoinn ns(NumDimensionsSynthetic, 30 /*age_max*/, 15 /*lambda*/, 0.1 /*C1*/, 0.05 /*C2*/); //Pathbased dataset mcl 1.2 20,30 итераций
nn::ESoinn ns(NumDimensionsSynthetic, 23 /*age_max*/, 25 /*lambda*/, 0.1 /*C1*/, 0.05 /*C2*/); //Compound dataset mcl 1.2 10,50 итераций
//nn::ESoinn ns(NumDimensionsSynthetic, 100 /*age_max*/, 50 /*lambda*/, 0.72 /*C1*/, 0.9 /*C2*/); //S2 dataset mcl 1.2 10,20 итераций
//nn::ESoinn ns(NumDimensionsSynthetic, 15 /*age_max*/, 30 /*lambda*/, 0.9 /*C1*/, 0.9 /*C2*/); //S4 dataset mcl 1.2 10,20 итераций
//----------------------------------------------
//nn::ESoinn ns(NumDimensionsSynthetic, 50 /*age_max*/, 60 /*lambda*/, 0.7 /*C1*/, 0.1 /*C2*/); //Jain dataset mcl 1.2
std::vector<std::vector<uint32_t>> conn_comp;
ns.trainNetwork(points, conn_comp, answers, 10, 50);
std::cout << "Number of clusters " << conn_comp.size() << std::endl;
for (uint32_t i = 0; i < conn_comp.size(); i++)
for (uint32_t j = 0; j < conn_comp[i].size(); j++)
neuron_clusters.emplace(conn_comp[i][j], i);
/*
ns.exportEdgesFile(output_filename);
//run mcl algorithm
std::string mcl_path = "/home/yura/local/bin/mcl ";
std::string output_mcl = "clusters";
std::string options = " -I 2.0 --abc -o ";
std::string command = mcl_path + output_filename + options + output_mcl;
system(command.c_str());
//read answer from mcl
readMCLAnswer(output_mcl);
*/
//print file for visualization
std::ofstream of("visualize_file", std::ios::out);
std::ofstream ofc("cluster_file", std::ios::out);
std::vector<uint32_t> find_clusters;
for (const auto p: points)
{
for (uint32_t i = 0; i < p->getNumDimensions(); i++)
{
of << p->getConcreteCoord(i);
ofc << p->getConcreteCoord(i) << " ";
if (i != p->getNumDimensions() - 1)
of << " ";
}
of << std::endl;
uint32_t cluster_id = ns.findPointCluster(p.get(), neuron_clusters);
ofc << cluster_id << std::endl;
find_clusters.push_back(cluster_id);
}
of.close();
ofc.close();
qm::QualityMeasures qlm(answers, find_clusters);
std::cout << "QUALITY MEASURES" << std::endl;
std::cout << "Purity = " << qlm.evalPurity() << std::endl;
std::cout << "Normalized mutual information = " << qlm.evalNormalizedMutualInformation() << std::endl;
qlm.findStatisticalQualityMeasures();
std::cout << "Rand index = " << qlm.randIndex() << std::endl;
std::cout << "precision = " << qlm.precision() << std::endl;
std::cout << "recall = " << qlm.recall() << std::endl;
std::cout << "Fscore(1) = " << qlm.Fscore(1) << std::endl;
//define clusters
/*
uint32_t i = 0;
for (const auto p: points)
{
log_netw->info((boost::format("%d ") % ns.findPointCluster(p.get(), neuron_clusters)).str(), true);
i++;
if (i % 50 == 0) log_netw->info("\n", true);
}
*/
ns.exportNetworkGDF("network.gdf", neuron_clusters);
log_netw->info("\n", true);
log_netw->info("The program is succesfully ended");
return 0;
}
