A=LOOM(H) returns a loom object initialized with hyperparameters H.
Hyperparameters, and their defaults
R=1 -- the soft margin parameter K=K-R*diag(diag(K))
A=0.001; -- additional regularization on alphas, A*||alpha||_1
ridge=1e-12; -- a ridge on the kernel
balanced_ridge=0 -- for unbalanced data
child=kernel -- the kernel is stored as a member called "child"
Model
alpha -- the weights
b0 -- the threshold
Xsv -- the Support Vectors
Methods:
train, test
|
Reference : Leave One Out Error, Stability, and Generalization of Voting Combinations of Classifiers |
|
Author : Theodoros Evgeniou ,Massimiliano Pontil , André Elisseeff |