void print_matrix(const real_2d_array& m, int row1, int row2, int col1, int col2) {
row1 = min(row1, m.rows());
row2 = min(row2, m.rows());
col1 = min(col1, m.cols());
col2 = min(col2, m.cols());
for (int r=row1; r<row2; ++r) {
for (int c=col1; c<col2; ++c) {
cout<<m[r][c]<<"\t";
}
cout<<endl;
}
}
void store_idata(const char* fpath, const real_2d_array& data) {
FILE* f = fopen(fpath, "w");
if (!f) throw "[store_data] Failed opening file!";
for (int row=0; row<data.rows(); ++row) {
for (int col=0; col<data.cols()-1; ++col) {
fprintf(f, "%i\t",(int)data[row][col]);
}
fprintf(f, "%i", (int)data[row][data.cols()-1]);
//if (row != data.rows()-1)
fprintf(f, "\n");
}
fclose(f);
}
CostCalculator_eigen::CostCalculator_eigen(const real_1d_array expectReturn,
const real_2d_array &varMatrix,
const real_1d_array &tradingCost,
const real_1d_array ¤tWeight) {
assert(expectReturn.length() == varMatrix.rows());
assert(varMatrix.rows() == varMatrix.cols());
assert(tradingCost.length() == varMatrix.rows());
assert(currentWeight.length() == varMatrix.rows());
variableNumber_ = expectReturn.length();
xReal_.resize(variableNumber_, 1);
varMatrix_.resize(variableNumber_, variableNumber_);
expectReturn_.resize(variableNumber_);
tradingCost_.resize(variableNumber_);
currentWeight_.resize(variableNumber_);
for (int i = 0; i != variableNumber_; ++i) {
expectReturn_(i) = expectReturn[i];
tradingCost_(i) = tradingCost[i];
currentWeight_(i) = currentWeight[i];
for (int j = 0; j != variableNumber_; ++j)
varMatrix_(i, j) = varMatrix[i][j];
}
}
// Construction of a kmean Trea for the neighborhoods stored in xy.
void kmeanTree(Node* node, real_2d_array xy, int min_size){
if(xy.rows()>min_size){ //Stop if we have less than min_size neighborhoods.
// K_mean (K=4) :
clusterizerstate s;
kmeansreport rep;
clusterizercreate(s);
clusterizersetpoints(s, xy, 2);
clusterizersetkmeanslimits(s, 5, 0);
clusterizerrunkmeans(s, 4, rep);
if(int(rep.terminationtype)==1){ // Check if the K_mean step finished correctly.
// Count the number of neighborhoods in each cluster.
vector<int> count(4);
for(int c=0;c<4;c++){
for(int i=0;i<rep.npoints;++i){
if(rep.cidx[i] == c){
count[c]++;
}
}
}
if(count[0] && count[1] && count[2] && count[3]){ // Check that each cluster is non-empty.
// Create new nodes
for(int c=0;c<4;c++){
Node* nodeSon = new Node;
node->addSon(nodeSon);
}
vector<real_2d_array> clusters(4);
for(int c=0;c<4;c++){
// Add the center element of each cluster in its related node.
for(int f=0;f<rep.nfeatures;++f){
node->getSon(c)->addCenterCoord((double)rep.c[c][f]);
}
// Separate the neighborhoods into the new nodes.
int id = 0;
clusters[c].setlength(count[c],rep.nfeatures);
for(int i=0;i<rep.npoints;++i){
if(rep.cidx[i] == c){
for(int f=0;f<rep.nfeatures;++f){
clusters[c][id][f] = xy[i][f];
}
id++;
}
}
}
// Iterate the construction on the new nodes.
for(int c=0;c<4;c++){
kmeanTree(node->getSon(c),clusters[c],min_size);
}
}
}
}
}