A=SVR(H) returns a svr object initialized with hyperparameters H.
Hyperparameters, and their defaults
C=Inf; -- the soft margin C parameter
optimizer='andre'; -- other choices={quadprog,andre,libsvm,svmtorch,sparse}
alpha_cutoff=-1; -- keep alphas with abs(a_i)>alpha_cutoff
nu = 0; -- nu parameter of a nu svr (different
from zero implies the nu-SVR is used
otherwise, the epsilon-SVR is used)
child=kernel; -- the kernel is stored as a member called "child"
epsilon=0.1; -- the value of epsilon in the epsilon
insensitive loss function during learning
use_signed_output=0; -- set to 1 implies that the svr is used
in classification (+1/-1 outputs), set to 0
implies that the svr is used in regression
Model
alpha -- the weights
b0 -- the threshold
Xsv -- the Support Vectors
Methods:
train, test, get_w
| |
Reference : A tutorial on Support Vector Regression |
| |
Author : Alex J. Smola , Bernhard Schölkopf |