def setOptimizer(self,
method,
num_restarts=None,
min_threshold=None,
meanRange=None,
covRange=None,
likRange=None):
'''
Overriding. Usage see base class pyGPs.gp.GP.setOptimizer
'''
conf = None
if (num_restarts != None) or (min_threshold != None):
conf = pyGPs.Optimization.conf.random_init_conf(
self.meanfunc, self.covfunc, self.likfunc)
conf.num_restarts = num_restarts
conf.min_threshold = min_threshold
if not meanRange is None:
conf.meanRange = meanRange
if not covRange is None:
conf.covRange = covRange
if not likRange is None:
conf.likRange = likRange
if method == "Minimize":
self.optimizer = opt.Minimize(self, conf)
elif method == "SCG":
self.optimizer = opt.SCG(self, conf)
elif method == "CG":
self.optimizer = opt.CG(self, conf)
elif method == "BFGS":
self.optimizer = opt.BFGS(self, conf)
def __init__(self):
super(GPC, self).__init__()
self.meanfunc = mean.Zero() # default prior mean
self.covfunc = cov.RBF() # default prior covariance
self.likfunc = lik.Erf() # erf likihood
self.inffunc = inf.EP() # default inference method
self.optimizer = opt.Minimize(self) # default optimizer
def __init__(self):
super(GPR, self).__init__()
self.meanfunc = mean.Zero() # default prior mean
self.covfunc = cov.RBF() # default prior covariance
self.likfunc = lik.Gauss() # likihood with default noise variance 0.1
self.inffunc = inf.Exact() # inference method
self.optimizer = opt.Minimize(self) # default optimizer
def setOptimizer(self,
method,
num_restarts=None,
min_threshold=None,
meanRange=None,
covRange=None,
likRange=None):
'''Set optimizer. See base class.'''
conf = None
if (num_restarts != None) or (min_threshold != None):
conf = pyGPs.Optimization.conf.random_init_conf(
self.meanfunc, self.covfunc, self.likfunc)
conf.num_restarts = num_restarts
conf.min_threshold = min_threshold
if meanRange != None:
conf.meanRange = meanRange
if covRange != None:
conf.covRange = covRange
if likRange != None:
conf.likRange = likRange
if method == "Minimize":
self.optimizer = opt.Minimize(self, conf)
elif method == "SCG":
self.optimizer = opt.SCG(self, conf)
elif method == "CG":
self.optimizer = opt.CG(self, conf)
elif method == "BFGS":
self.optimizer = opt.BFGS(self, conf)
elif method == "LBFGSB":
self.optimizer = opt.LBFGSB(self, conf)
elif method == "COBYLA":
self.optimizer = opt.COBYLA(self, conf)
def __init__(self):
super(GPC_FITC, self).__init__()
self.meanfunc = mean.Zero() # default prior mean
self.covfunc = cov.RBF() # default prior covariance
self.likfunc = lik.Erf() # erf liklihood
self.inffunc = inf.FITC_EP() # default inference method
self.optimizer = opt.Minimize(self) # default optimizer
self.u = None # no default inducing points
