def wsabil_fixed(base_gp_data):
base_gp, X, Y = base_gp_data
wsabil = WSABIL(base_gp=base_gp,
X=base_gp.X,
Y=base_gp.Y,
adapt_alpha=True)
return wsabil, X, Y
bound = np.min(base_gp.Y) - 0.5
model = BoundedBayesianQuadrature(base_gp=base_gp,
X=X,
Y=Y,
bound=bound,
is_lower_bounded=True)
elif METHOD == "Bounded BQ upper":
bound = np.max(base_gp.Y) + 0.5
model = BoundedBayesianQuadrature(base_gp=base_gp,
X=X,
Y=Y,
bound=bound,
is_lower_bounded=False)
elif METHOD == "WSABI-l adapt":
model = WSABIL(base_gp=base_gp,
X=base_gp.X,
Y=base_gp.Y,
adapt_alpha=True)
elif METHOD == "WSABI-l fixed":
model = WSABIL(base_gp=base_gp,
X=base_gp.X,
Y=base_gp.Y,
adapt_alpha=False)
else:
raise ValueError
print()
print("method: {}".format(METHOD))
print("no dimensions: {}".format(D))
print()
def wsabil_adapt(base_gp_data):
base_gp, X, Y = base_gp_data
wsabil = WSABIL(base_gp=base_gp, X=X, Y=Y, adapt_alpha=True)
return wsabil, X, Y
def wsabil_fixed(base_gp):
wsabil = WSABIL(base_gp=base_gp,
X=base_gp.X,
Y=base_gp.Y,
adapt_alpha=False)
return wsabil, None
def wsabil_adapt(base_gp):
wsabil = WSABIL(base_gp=base_gp,
X=base_gp.X,
Y=base_gp.Y,
adapt_alpha=True)
return wsabil, None
def get_wsabil_fixed():
base_gp, dat = get_base_gp()
wsabil = WSABIL(base_gp=base_gp, X=dat.X, Y=dat.Y, adapt_alpha=False)
return wsabil
def get_wsabil_adapt():
base_gp, dat = get_base_gp()
return WSABIL(base_gp=base_gp, X=dat.X, Y=dat.Y, adapt_alpha=True)