wolffd@0: wolffd@0: wolffd@0: wolffd@0: Netlab Reference Manual som wolffd@0: wolffd@0: wolffd@0: wolffd@0:

som wolffd@0:

wolffd@0:

wolffd@0: Purpose wolffd@0:

wolffd@0: Creates a Self-Organising Map. wolffd@0: wolffd@0:

wolffd@0: Synopsis wolffd@0:

wolffd@0:

wolffd@0: 
wolffd@0: net = som(nin, map_size)
wolffd@0:

wolffd@0: wolffd@0: wolffd@0:

wolffd@0: Description wolffd@0:

wolffd@0: net = som(nin, map_size) creates a SOM net wolffd@0: with input dimension (i.e. data dimension) nin and map dimensions wolffd@0: map_size. Only two-dimensional maps are currently implemented. wolffd@0: wolffd@0:

The fields in net are wolffd@0:

wolffd@0: 
wolffd@0:   type = 'som'
wolffd@0:   nin = number of inputs
wolffd@0:   map_dim = dimension of map (constrained to be 2)
wolffd@0:   map_size = grid size: number of nodes in each dimension
wolffd@0:   num_nodes = number of nodes: the product of values in map_size
wolffd@0:   map = map_dim+1 dimensional array containing nodes
wolffd@0:   inode_dist = map of inter-node distances using Manhatten metric
wolffd@0:

wolffd@0: wolffd@0: wolffd@0:

The map contains the node vectors arranged column-wise in the first wolffd@0: dimension of the array. wolffd@0: wolffd@0:

wolffd@0: See Also wolffd@0:

wolffd@0: kmeans, somfwd, somtrain

wolffd@0: Pages: wolffd@0: Index wolffd@0:

wolffd@0: