void addMatrixRow(MatDoub U, int row, MatDoub &out) {
int dummy = -1000;
for(int i=0; i<out.nrows(); i++) {
out[i][row] = dummy;
out[row][i] = dummy;
}
datain = U.getMatrixArray();
datainO = out.getMatrixArray();
int k = 0;
data [out.nrows()*out.nrows()];
for(int i=0; i < out.nrows()*out.ncols(); i++) {
if( datainO[i] == dummy )
data[i] = 0;
else
data[i] = datain[k++];
}
out = MatDoub( out.nrows(), out.nrows(), data );
}
void Eigsym::eig(const MatDoub &A, MatDoub &V, VecDoub &lambda) {
unsigned int n = A.ncols(); /* size of the matrix */
double a[n*n]; /* store initial matrix */
double w[n]; /* store eigenvalues */
int matz = 1; /* return both eigenvalues as well as eigenvectors */
double x[n*n]; /* store eigenvectors */
for(unsigned int i=0; i<n; i++) {
for(unsigned int j=0; j<n; j++) {
a[i*n+j] = A[i][j];
}
}
unsigned int ierr = 0;
ierr = rs ( n, a, w, matz, x );
V.assign(n,n,0.0);
lambda.resize(n);
for(unsigned int i=0; i<n; i++) {
lambda[i] = w[i];
for(unsigned int j=0; j<n; j++) {
V[j][i] = x[i*n+j];
}
}
}
void dump_nrmat( MatDoub &m ) {
for( int r=0; r<m.nrows(); r++ ) {
for( int c=0; c<m.ncols(); c++ ) {
printf( "%+3.2le ", m[r][c] );
}
printf( "\n" );
}
}
/*
Calculte the Modularity matrix when split
into more than two communities, see [2]
in method declarations above.
*/
void calculateB(MatDoub B, MatDoub &Bg) {
int Ng = B.ncols();
if( Bg.ncols() != Ng )
Bg.resize(Ng,Ng);
for(int i=0; i<Ng; i++) {
for(int j=0; j<Ng; j++) {
double sum = 0.0;
for(int k=0; k<Ng; k++)
sum += B[i][k];
Bg[i][j] = B[i][j] -1.0 * delta(i,j) * sum;
}
}
}
void copyNRMatToZMat( MatDoub &m, ZMat &z ) {
// account for NR3 is rowmajor, ZMat is colmajor.
int rows = m.nrows();
int cols = m.ncols();
if( z.rows != rows || z.cols != cols ) {
z.alloc( rows, cols, zmatF64 );
}
for( int r=0; r<rows; r++ ) {
for( int c=0; c<cols; c++ ) {
z.putD( r, c, m[r][c] );
}
}
}
/*
Calculate the eigenvalues, betai, and eigenvectors, u, for
the current Modularity matrix Bgi.
*/
void calculateEigenVectors() {
int Ng = Bgi.ncols();
if(u.ncols() != Ng) {
u.resize(Ng,Ng);
betai.resize(Ng);
}
u.resize(Ng,Ng);
betai.resize(Ng);
Symmeig h(Bgi, true);
for(int i=0; i<Ng; i++) {
betai[i] = h.d[i];
for(int j=0; j<Ng; j++) {
u[j][i] = h.z[j][i];
}
}
}
void removeMatrixRow( MatDoub &out ) {
int dummy = -1000;
datain = Ri.getMatrixArray();
data [Ri.nrows()*Ri.nrows()];
int k=0;
for(int i=0; i < Ri.nrows()*Ri.ncols(); i++) {
double ele = datain[i];
if(ele != dummy)
data[k++] = ele;
}
out = MatDoub( out.nrows(), out.nrows(), data );
}
/*
Find the leading eigenvector, i.e.
the one which corresponds to the most positive
eigenvalue.
*/
void findLeadingEigenVectors(int &ind) {
int Ng = Bgi.ncols();
int ind_max = 0;
int ind_min = 0;
double max = betai[ind_max];
double min = betai[ind_min];
for(int i=0; i<Ng-1; i++) {
if( betai[i] > max ) {
max = betai[i];
ind_max = i;
}
}
ind = ind_max;
}
/*
Utility method used by RandomWalk algorithm to
resize the Graph Laplacian for each community, com,
within the network.
*/
void getSubMatrix(int com, vector<node> &Nodes) {
int dummy = -1000;
int rows = 0;
Rh.resize(R.nrows(), R.nrows());
Rh = R;
//--- NR style
for( int i=0; i< C.size(); i++) {
if( C[i] == com )
Nodes.push_back(node(rows++,0.0,0.0));
else {
for( int k=0; k<Rh.nrows(); k++) {
Rh[i][k] = dummy;
Rh[k][i] = dummy;
}
}
}
datain = Rh.getMatrixArray();
data [Rh.nrows()*Rh.nrows()];
int ind = 0;
for(int i=0; i < Rh.nrows()*Rh.ncols(); i++) {
double ele = datain[i];
if(ele != dummy)
data[ind++] = ele;
}
Ri.resize(rows,rows);
Ri = MatDoub( rows, rows, data );
}
/*
Update the index vectors, si and SI, for each node in the
current split such that:
si(i) = 1 if eigenvector_max(i) > 0
= -1 if eigenvector_max(i) < 0
SI(i,0) = 1
SI(i,1) = 0 if eigenvector_max(i) > 0
= 0
= 1 if eigenvector_max(i) < 0
*/
void maximiseIndexVectors( int ind ) {
int Ng = u.ncols();
si.resize(Ng);
SI.resize(Ng,2);
for(int i=0; i<Ng; i++) {
if(u[i][ind] < 0) {
si[i] = -1;
SI[i][0] = 0;
SI[i][1] = 1;
} else {
si[i] = 1;
SI[i][0] = 1;
SI[i][1] = 0;
}
}
}
