void example_som(int nrows, int ncols, double** data, int** mask)
/* Calculate a self-organizing map, applied to genes */
{ int i, j, k;
const int nxgrid = 2;
const int nygrid = 2; /* Rectangular grid 2x2 */
const double inittau = 0.02; /* Initial value of the neighborhood function */
const int niter = 1000; /* Number of iterations */
const char dist = 'c'; /* Pearson correlation */
double* weight = malloc(ncols*sizeof(double));
int (*clusterid)[2] = malloc(nrows*sizeof(int[2]));
double*** celldata = malloc(nxgrid*sizeof(double**));
for (i = 0; i < nxgrid; i++)
{ celldata[i] = malloc(nygrid*sizeof(double*));
for (j = 0; j < nygrid; j++)
celldata[i][j] = malloc(ncols*sizeof(double));
}
for (i = 0; i < ncols; i++) weight[i] = 1.0;
printf("======================= Self-Organizing Map ===================\n");
printf("(results may change on every run)\n");
somcluster(nrows, ncols, data, mask, weight, 0, nxgrid, nygrid, inittau,
niter, dist, celldata, clusterid);
printf("Cluster assignments:\n");
for(i=0; i<nrows; i++)
printf("Gene %2d: %2d %2d\n",i,clusterid[i][0],clusterid[i][1]);
printf("Cluster centroids:\n");
printf("\t");
for (j = 0; j < ncols; j++) printf("\tCol %d", j);
printf ("\n");
for (i = 0; i < nxgrid; i++)
{ for (j = 0; j < nygrid; j++)
{ printf("Cell (%d,%d):", i, j);
for (k = 0; k < ncols; k++) printf("\t%5.2g",celldata[i][j][k]);
printf("\n");
}
}
printf("\n");
/* Deallocate memory */
for (i = 0; i < nxgrid; i++)
{ for (j = 0; j < nygrid; j++) free(celldata[i][j]);
free(celldata[i]);
}
free(celldata);
free(weight);
free(clusterid);
return;
}
static int
PerformArraySOM(FILE* file, int XDim, int YDim, int iterations, double tau,
char metric)
{ int i = 0;
int j = 0;
int k;
int ok;
int (*Group)[2] = malloc(_columns*sizeof(int[2]));
double*** Nodes = malloc(XDim*sizeof(double**));
int* clusterid = malloc(_columns*sizeof(int));
if (Nodes)
{ for (i = 0; i < XDim; i++)
{ Nodes[i] = malloc(YDim*sizeof(double*));
j = 0;
if (!Nodes[i]) break;
for ( ; j < YDim; j++)
{ Nodes[i][j] = malloc(_rows*sizeof(double));
if (!Nodes[i][j]) break;
}
if (j < YDim) break;
}
}
if (!Group || !clusterid || !Nodes || i < XDim || j < YDim)
{ if (Group) free(Group);
if (Nodes)
{ if (i < XDim)
{ while (j--) free(Nodes[i][j]);
free(Nodes[i]);
}
while (i--)
{ for (j = 0; j < YDim; j++) free(Nodes[i][j]);
free(Nodes[i]);
}
free(Nodes);
}
free(clusterid);
return 0;
}
somcluster(_rows, _columns, _data, _mask, _geneweight, 1,
XDim, YDim, tau, iterations, metric, Nodes, Group);
for (i=0; i<_columns; i++)
clusterid[i] = Group[i][0] * YDim + Group[i][1];
free(Group);
fprintf(file, "%s\t", _uniqID);
for (i=0; i<XDim; i++)
for (j=0; j<YDim; j++) fprintf(file, "\tNODE(%d,%d)", i, j);
putc('\n', file);
for (k=0;k<_rows;k++)
{ int index = _geneindex[k];
fprintf(file, "%s\t", _geneuniqID[index]);
if (_genename[index]) fputs(_genename[index], file);
else fputs(_geneuniqID[index], file);
for (i=0; i<XDim; i++)
for (j=0; j<YDim; j++) fprintf(file, "\t%f", Nodes[i][j][index]);
putc('\n', file);
}
for (i=0;i<XDim;i++)
{ for (j=0; j<YDim; j++) free(Nodes[i][j]);
free(Nodes[i]);
}
free(Nodes);
ok = SetClusterIndex('a', XDim * YDim, clusterid);
free(clusterid);
return ok;
}
static int
PerformGeneSOM(FILE* file, int XDim, int YDim, int iterations, double tau,
char metric)
{ int i = 0;
int j = 0;
int k;
int ok;
int (*Group)[2] = malloc(_rows*sizeof(int[2]));
double*** Nodes = malloc(XDim*sizeof(double**));
int* clusterid = malloc(_rows*sizeof(int));
int* index = malloc(_columns*sizeof(int));
if (Nodes)
{ for (i = 0; i < XDim; i++)
{ Nodes[i] = malloc(YDim*sizeof(double*));
j = 0;
if (!Nodes[i]) break;
for ( ; j < YDim; j++)
{ Nodes[i][j] = malloc(_columns*sizeof(double));
if (!Nodes[i][j]) break;
}
if (j < YDim) break;
}
}
if (!Group || !clusterid || !index || !Nodes || i < XDim || j < YDim)
{ if (Group) free(Group);
if (clusterid) free(clusterid);
if (index) free(index);
if (Nodes)
{ if (i < XDim)
{ while (j--) free(Nodes[i][j]);
free(Nodes[i]);
}
while (i--)
{ for (j = 0; j < YDim; j++) free(Nodes[i][j]);
free(Nodes[i]);
}
free(Nodes);
}
return 0;
}
somcluster(_rows, _columns, _data, _mask, _arrayweight, 0,
XDim, YDim, tau, iterations, metric, Nodes, Group);
for (i=0; i<_rows; i++) clusterid[i] = Group[i][0] * YDim + Group[i][1];
free(Group);
for (k=0; k<_columns; k++) index[k] = k;
sort(_columns, _arrayorder, index);
fputs("NODE", file);
for (i=0; i<_columns; i++) fprintf(file, "\t%s", _arrayname[index[i]]);
putc('\n', file);
for (i=0; i<XDim; i++)
{ for (j=0; j<YDim; j++)
{ fprintf(file, "NODE(%d,%d)", i, j);
for (k=0; k<_columns; k++) fprintf(file, "\t%f", Nodes[i][j][index[k]]);
putc('\n', file);
}
}
free(index);
for (i=0;i<XDim;i++)
{ for (j=0; j<YDim; j++) free(Nodes[i][j]);
free(Nodes[i]);
}
free(Nodes);
ok = SetClusterIndex('g', XDim * YDim, clusterid);
free(clusterid);
return ok;
}
/* @api private */
VALUE rb_do_self_organizing_map(int argc, VALUE *argv, VALUE self) {
VALUE nx, ny, data, mask, weights, options;
rb_scan_args(argc, argv, "31", &nx, &ny, &data, &options);
if (TYPE(data) != T_ARRAY)
rb_raise(rb_eArgError, "data should be an array of arrays");
mask = get_value_option(options, "mask", Qnil);
if (!NIL_P(mask) && TYPE(mask) != T_ARRAY)
rb_raise(rb_eArgError, "mask should be an array of arrays");
if (NIL_P(nx) || NUM2INT(rb_Integer(nx)) <= 0)
rb_raise(rb_eArgError, "nx should be > 0");
if (NIL_P(ny) || NUM2INT(rb_Integer(ny)) <= 0)
rb_raise(rb_eArgError, "ny should be > 0");
int nxgrid = NUM2INT(rb_Integer(nx));
int nygrid = NUM2INT(rb_Integer(ny));
int transpose = get_int_option(options, "transpose", 0);
int npass = get_int_option(options, "iterations", DEFAULT_ITERATIONS);
// e = euclidian,
// b = city-block distance
// c = correlation
// a = absolute value of the correlation
// u = uncentered correlation
// x = absolute uncentered correlation
// s = spearman's rank correlation
// k = kendall's tau
int dist = get_int_option(options, "metric", 'e');
double tau = get_dbl_option(options, "tau", 1.0);
int i, j, k;
int nrows = RARRAY_LEN(data);
int ncols = RARRAY_LEN(rb_ary_entry(data, 0));
double **cdata = (double**)malloc(sizeof(double*)*nrows);
int **cmask = (int **)malloc(sizeof(int *)*nrows);
double *cweights = (double *)malloc(sizeof(double )*ncols);
int **ccluster;
double ***ccelldata;
int dimx = nrows, dimy = ncols;
if (transpose) {
dimx = ncols;
dimy = nrows;
}
ccluster = (int **)malloc(sizeof(int*)*dimx);
for (i = 0; i < dimx; i++)
ccluster[i] = (int*)malloc(sizeof(int)*2);
for (i = 0; i < nrows; i++) {
cdata[i] = (double*)malloc(sizeof(double)*ncols);
cmask[i] = (int *)malloc(sizeof(int )*ncols);
for (j = 0; j < ncols; j++) {
cdata[i][j] = NUM2DBL(rb_Float(rb_ary_entry(rb_ary_entry(data, i), j)));
cmask[i][j] = NIL_P(mask) ? 1 : NUM2INT(rb_Integer(rb_ary_entry(rb_ary_entry(mask, i), j)));
}
}
weights = NIL_P(options) ? Qnil : rb_hash_aref(options, ID2SYM(rb_intern("weights")));
for (i = 0; i < ncols; i++) {
cweights[i] = NIL_P(weights) ? 1.0 : NUM2DBL(rb_Float(rb_ary_entry(weights, i)));
}
ccelldata = (double***)malloc(sizeof(double**)*nxgrid);
for (i = 0; i < nxgrid; i++) {
ccelldata[i] = (double **)malloc(sizeof(double*)*nygrid);
for (j = 0; j < nygrid; j++)
ccelldata[i][j] = (double *)malloc(sizeof(double)*dimy);
}
somcluster(nrows, ncols, cdata, cmask, cweights, transpose, nxgrid, nygrid, tau, npass, dist, ccelldata, ccluster);
VALUE result = rb_hash_new();
VALUE cluster = rb_ary_new();
VALUE centroid = rb_ary_new();
for (i = 0; i < dimx; i++) {
VALUE gridpoint = rb_ary_new();
rb_ary_push(gridpoint, INT2NUM(ccluster[i][0]));
rb_ary_push(gridpoint, INT2NUM(ccluster[i][1]));
rb_ary_push(cluster, gridpoint);
}
for (i = 0; i < nxgrid; i++) {
for (j = 0; j < nygrid; j++) {
VALUE point = rb_ary_new();
for (k = 0; k < dimy; k++)
rb_ary_push(point, DBL2NUM(ccelldata[i][j][k]));
rb_ary_push(centroid, point);
}
}
rb_hash_aset(result, ID2SYM(rb_intern("cluster")), cluster);
rb_hash_aset(result, ID2SYM(rb_intern("centroid")), centroid);
for (i = 0; i < nrows; i++) {
free(cdata[i]);
free(cmask[i]);
}
for (i = 0; i < dimx; i++)
free(ccluster[i]);
for (i = 0; i < nxgrid; i++) {
for (j = 0; j < nygrid; j++)
free(ccelldata[i][j]);
free(ccelldata[i]);
}
free(cdata);
free(cmask);
free(ccelldata);
free(cweights);
free(ccluster);
return result;
}
