void example_distance_array(int nrows, int ncols, double** data, int** mask)
/* Calculate the distance matrix between microarrays using the Spearman rank
* correlation. */
{ int i, j;
double** distMatrix;
double* weight = malloc(nrows*sizeof(double));
printf("========== Spearman distance matrix between microarrays =========\n");
for (i = 0; i < nrows; i++) weight[i] = 1.0;
distMatrix = distancematrix(nrows, ncols, data, mask, weight, 's', 1);
if (!distMatrix)
{ printf ("Insufficient memory to store the distance matrix\n");
free(weight);
return;
}
printf(" Microarray:");
for(i=0; i<ncols-1; i++) printf("%9d", i);
printf("\n");
for(i=0; i<ncols; ++i)
{ printf("Microarray %2d: ",i);
for(j=0; j<i; ++j) printf(" %f",distMatrix[i][j]);
printf("\n");
}
printf("\n");
free(weight);
for(i = 0; i < ncols; i++) free(distMatrix[i]);
free(distMatrix);
return;
}
void HClusterDlg::SpatialConstraintClustering()
{
int transpose = 0; // row wise
fastcluster::cluster_result Z2(rows-1);
if (chk_contiguity->GetValue()) {
vector<boost::uuids::uuid> weights_ids;
WeightsManInterface* w_man_int = project->GetWManInt();
w_man_int->GetIds(weights_ids);
int sel = combo_weights->GetSelection();
if (sel < 0) sel = 0;
if (sel >= weights_ids.size()) sel = weights_ids.size()-1;
boost::uuids::uuid w_id = weights_ids[sel];
GalWeight* gw = w_man_int->GetGal(w_id);
if (gw == NULL) {
wxMessageDialog dlg (this, _("Invalid Weights Information:\n\n The selected weights file is not valid.\n Please choose another weights file, or use Tools > Weights > Weights Manager\n to define a valid weights file."), _("Warning"), wxOK | wxICON_WARNING);
dlg.ShowModal();
return;
}
//GeoDaWeight* weights = w_man_int->GetGal(w_id);
// Check connectivity
if (!CheckConnectivity(gw)) {
wxString msg = _("The connectivity of selected spatial weights is incomplete, please adjust the spatial weights.");
wxMessageDialog dlg(this, msg, _("Warning"), wxOK | wxICON_WARNING );
dlg.ShowModal();
return;
}
double** ragged_distances = distancematrix(rows, columns, input_data, mask, weight, dist, transpose);
double** distances = DataUtils::fullRaggedMatrix(ragged_distances, rows, rows);
for (int i = 1; i < rows; i++) free(ragged_distances[i]);
free(ragged_distances);
std::vector<bool> undefs(rows, false);
SpanningTreeClustering::AbstractClusterFactory* redcap = new SpanningTreeClustering::FirstOrderSLKRedCap(rows, columns, distances, input_data, undefs, gw->gal, NULL, 0);
for (int i=0; i<redcap->ordered_edges.size(); i++) {
Z2[i]->node1 = redcap->ordered_edges[i]->orig->id;
Z2[i]->node2 = redcap->ordered_edges[i]->dest->id;
Z2[i]->dist = redcap->ordered_edges[i]->length;
}
delete redcap;
}
}
bool BagOfFeatures::buildHierarchicalTree(int transpose,
char dist,
char method,
double** distmatrix)
{
//double** buildMatrixHierarchical(const ImageFeatures *featurePts, const int feature_count,
// const int image_count, Node*& tree )
// Check to makes sure that features were found
if(trainObject == NULL || numFeatures == 0 || descrSize == 0)
return false;
int i, j;
int k = 0, l = 0, m;
int size;
int totalImages = 0;
cout << "Initializing the data..." << endl;
// Initialize the data
hClusterData = new double* [numFeatures];
for(i = 0; i < numFeatures; i++)
hClusterData[i] = new double [descrSize];
// Allocate mask and set it all to 1 (assume no missing data)
int ** mask = new int* [numFeatures];
for(i = 0; i < numFeatures; i++)
{
mask[i] = new int [descrSize];
for(j = 0; j < descrSize; j++)
mask[i][j] = 1;
}
// Set the weights equal, all 1
double* weight = new double [descrSize];
for(i = 0; i < descrSize; i++)
weight[i] = 1.0;
// For each class
for(m = 0; m < numClasses; m++)
{
totalImages = data[m].getTrainSize();
// For each image in that class...
for(l = 0; l < totalImages; l++)
{
size = trainObject[m].featureSet[l].size;
// for each feature in that image...
for(i = 0; i < size; i++)
{
// Copy the descriptor into the data array
for(j = 0; j < descrSize; j++)
{
hClusterData[k][j] = (double)trainObject[m].featureSet[l].descriptors[i][j];
//cout << hClusterData[k][j] << " ";
}
//cout << endl;
k++;
}
}
}
cout << "Calculating the distance matrix..." << endl;
distmatrix = distancematrix (descrSize, numFeatures, hClusterData, mask, weight, dist, transpose);
cout << "Building Hierarchical Tree" << endl;
// Centroid Hierarchical Clustering
// feature_count X DESCRIPTOR_SIZE
// The feature vectors
// mask (all 1s)
// weights (all 1s)
hTree = treecluster(descrSize, numFeatures, hClusterData, mask, weight, transpose, dist, method, distmatrix);
// Release the mask
for(i = 0; i < numFeatures; i++)
delete [] mask[i];
delete [] mask;
// Release the weight
delete [] weight;
// Make sure that the tree was allocated
if(!hTree)
{
cout << "Could not allocate the tree: Insufficient memory..." << endl;
for(i = 0; i < numFeatures; i++)
delete [] hClusterData[i];
delete [] hClusterData;
return false;
}
return true;
}
