| |
TOY data for Regression problems
|
A=toyreg(H) returns a toy regression object initialized with hyperparameters H.
This object can generate toy data for regression with
N input dimensions and O output dimensions.
If "nonlinearity" is specified it uses a map object to calculate the
outputs.
Hyperparameters, and their defaults
l=100 -- examples
noiselevel -- 0
n=1 -- input dimension
o=1 -- output dimension (possibly overriden by outpmap)
nonlinear=0 -- if 1 uses "outpmap" to calculate Y=map(X)
outpmap=map;
seed=-1 -- random seed used to generate it, if -1 do not
set seed
Methods:
generate,train,test