def __init__(self,
X,
y,
bounds,
xbounds,
initial_guess,
l_INI,
sample_size=10000,
num_ini_guess=2,
alpha=10.):
self.n = X.shape[1]
self.sigma_inv = np.ones(self.n) # default value
self.model = Kriging(self.sigma_inv, xbounds, num_ini_guess,
sample_size) # Fit Kriging on the data.
self.model.fit(X, y)
self.input = X
self.rem_eng = y
self.sct = None
self.alpha = alpha
self.l_INI = l_INI
self.initial_guess = initial_guess
self.bounds = bounds
import scipy.optimize as opt
all_sigs = np.zeros((len(pre.full_tab['id'].tolist()), 31))
all_improv = np.zeros_like(pre.full_tab['id'].tolist())
lb = 0.01
ub = 100.
bounds = [(lb, ub)] * 31
for n, i in enumerate(pre.full_tab['id'].tolist()):
a, b = pre.prep_by_id(i)
print i, len(a)
if len(a) == 1 or len(a) == 0:
all_improv[n] = 0
continue
krig = Kriging(unit_sig)
krig.fit(a, b)
x0 = np.ones(31)
func = lambda x: -krig.f_by_sig(x)
res = opt.minimize(func,
x0=x0,
bounds=bounds,
method='SLSQP',
tol=1e-5,
options={
'eps': 1e-2,
'iprint': 2,
'disp': True,
'maxiter': 100
})
# all_improv[n] = np.nan_to_num(krig.f(a[-1])[0])
