def plot_rbf(data, n_center, theta_opt, n_line_precision=100):
"""
Create of plot that shows the RBF expansion and the fit as compared to the scattered data sets.
:param data:
:param n_center:
:param theta_opt:
:param n_line_precision:
:return:
"""
fig, ax_list = plt.subplots(2, 2)
plt.subplots_adjust(hspace=0.4)
# Plot the polynomials
xx = np.linspace(-1, 1, n_line_precision).reshape((n_line_precision, 1))
centers, sigma = rbf.get_centers_and_sigma(n_center)
XX = rbf.design_matrix(xx, centers, sigma)
ax_list[0, 0].plot(xx, XX, linewidth=3)
ax_list[0, 0].set_xlabel('x')
ax_list[0, 0].set_ylabel('y')
ax_list[0, 0].set_xlim([-1, 1])
ax_list[0, 0].set_ylim([0, 1])
ax_list[0, 0].set_title('{} RBF kernels'.format(n_center))
# Computation the predicted model
y_pred = XX.dot(theta_opt)
# Plot the fit on training data
As = [(0, 1), (1, 0), (1, 1)]
Xs = ['x_train', 'x_val', 'x_test']
Ys = ['y_train', 'y_val', 'y_test']
Cs = ['blue', 'red', 'purple']
Titles = ['train', 'validation', 'test']
for a, x, y, c, ti in zip(As, Xs, Ys, Cs, Titles):
# Plot
ax_list[a].plot(xx, y_pred, color='black', linewidth=3)
ax_list[a].scatter(data[x], data[y], color=c, label=ti + ' set')
ax_list[a].set_xlabel('x')
ax_list[a].set_ylabel('y')
mse = rbf.compute_error(theta_opt, n_center, data[x], data[y])
mse = round(float(mse), 2)
#ax_list[a].set_title('Set {} (MSE {:.3g}) '.format(ti, mse))
ax_list[a].set_title('Set {} (MSE {}) '.format(ti, mse))
ax_list[a].set_xlim([-1, 1])
ax_list[a].set_ylim([-5, 5])
import numpy as np
import rbf
#x = np.array([[1, 2, 3], [4, 5, 6]])
#a = x.shape
#b = x.shape[1]
x = np.arange(0, 5, 1)
degree = 4
N = x.shape[0]
X = np.zeros((N, degree + 1))
for j in range(0, degree + 1): #loop from 0 to degree+1 -1 pay attention
powerj = np.power(x, j)
X[:, j] = powerj
print(X)
centers, sigma = rbf.get_centers_and_sigma(2)
design