def test_gradient(X,y):
tired = False
eps = 0.01
counter = 0
d = X.shape[1]
while not tired and counter < 100:
counter = counter + 1
phi=np.random.uniform(size=(d,1))
log_grad_formula = log_grad(X,y,phi)
log_grad_formula_norm = np.sqrt(np.sum(log_grad_formula**2))
log_grad_numerical = (log_cost(X,y,phi+eps) - log_cost(X,y,phi-eps))/(2*eps*sqrt(d))
log_grad_numerical_norm = np.sqrt(np.sum(log_grad_numerical**2))
diff = np.sqrt(np.sum((log_grad_formula-log_grad_numerical)**2))
print("||log_grad(phi)|| = %s"%log_grad_formula_norm)
print("||approximated|| = %s"%log_grad_numerical_norm)
print("||diff|| = %s"%(diff))
def test_cost(X,y):
tired = False
eps = 0.01
counter = 0
while not tired and counter < 20:
counter = counter + 1
phi=np.random.uniform(size=(X.shape[1],1))
print(log_cost(X, y, phi))