BPROJ Backward Projection using the SOM [pr,res,bmu] = bproj(g,smap,projtype) g: (Rout x Q) vector in the output space with data points by columns smap: SOM struct (see INISOM, BSOM) projtype: String with the projection type: 'discrete: Original discrete SOM projection (default) 'grnn': Generalized Regression Neural Network 'ls': Uses seast squares (Ridge Regression) to compute the weight matrix CALLS [pr,res,bmu] = bproj(p,smap) [pr,res,bmu] = bproj(p,smap,projtype) DESCRIPTION Projects output space samples (given as columns of data matrix g) onto the input space using the projection defined by the som 'smap'. Projection can be either discrete or continuous. See also FPROJ