# my_GP = GaussianProcess(SMKernel(X.shape[0], 3), X, LogisticFunction())
y = np.matrix(
[-1]*N + [1]*N + [-1]*N
).T
my_GP.add_task(y)
y2 = []
for i in range(3*N):
if X[0,i] > 0:
y2.append(1)
else:
y2.append(-1)
y2 = np.matrix(y2).T
my_GP.add_task(y2)
# print [-1+0.1*i for i in range(30)]
# x_star = np.matrix([-1+0.1*i for i in range(50)])
x_star = np.matrix(dnorm(0,sigma,N))
# f_mode, log_ML = my_GP.gpc_find_mode([0, 1], my_GP.cov_function.INITIAL_GUESS)
# print my_GP.gpc_optimize([0])
tasks = [0]
my_GP.hyperparameters = my_GP.gpc_optimize(tasks)[0]
print my_GP.hyperparameters
y_star = my_GP.gpc_make_prediction([0], x_star)
for i in range(len(y_star)):
print x_star[0, i], y_star[i]