def test_bayes_evidence(self):
k = pb.KalmanFilter(**self.setup_2)
y = np.array([[4.1], [-0.2], [1.4], [-2.1]])
u = np.array([[4.8], [-0.3], [1.1], [-1.8]])
exp_mu = np.array([
[ 3.62004716, -0.46320771],
[-0.16638519, 3.58787721],
[ 1.21108425, 0.0224309 ],
[-1.87141692, 0.98517451]
])
exp_var = np.array([
[ 0.00999960, 40.3342872],
[ 0.00999029, 0.20999610],
[ 0.00963301, 0.20962422],
[ 0.00963191, 0.20930431]
])
# in: y_t -1, y_t, y_t + 1
exp_evidences_log = np.array([
[ -3.68958564, -3.68287128, -3.68015356],
[ -3.16749303, -2.75744797, -2.44453198],
[ -2.69696927, -8.75357053, -18.48011903],
[-10.17228316, -3.47352413, -0.45566728]
])
for i in range(4):
k.bayes(y[i], u[i])
post = k.posterior()
self.assertApproxEqual(post.mean(), exp_mu[i])
self.assertApproxEqual(post.variance(), exp_var[i])
evidences = np.array([k.evidence_log(y[i] - 1.), k.evidence_log(y[i]), k.evidence_log(y[i] + 1.)])
self.assertApproxEqual(evidences, exp_evidences_log[i])
def test_bayes(self):
k = pb.KalmanFilter(**self.setup_2)
y = np.array([[4.1], [-0.2], [1.4], [-2.1]])
u = np.array([[4.8], [-0.3], [1.1], [-1.8]])
exp_mu = np.array([[3.62004716, -0.46320771],
[-0.16638519, 3.58787721], [1.21108425, 0.0224309],
[-1.87141692, 0.98517451]])
for i in range(4):
k.bayes(y[i], u[i])
mu = k.posterior().mu
self.assertApproxEqual(mu, exp_mu[i])
def run_kalman_on_mat_data(input_file, output_file, timer):
# this should be here so that only this stress fails when scipy is not installed
loadmat = scipy.io.loadmat
savemat = scipy.io.savemat
d = loadmat(input_file, struct_as_record=True, mat_dtype=True)
mu0 = np.reshape(d.pop('mu0'),
(-1, )) # otherwise we would get 2D array of shape (1xN)
P0 = d.pop('P0')
y = d.pop('y').T
u = d.pop('u').T
#x = d.pop('x').T
x = None
gauss = pb.GaussPdf(mu0, P0)
kalman = pb.KalmanFilter(d['A'], d['B'], d['C'], d['D'], d['Q'], d['R'],
gauss)
N = y.shape[0]
n = mu0.shape[0]
mean = np.zeros((N, n))
var = np.zeros((N, n))
timer.start()
for t in range(1, N): # the 1 start offset is intentional
kalman.bayes(y[t], u[t])
mean[t] = kalman.posterior().mean()
#var[t] = kalman.posterior().variance()
timer.stop()
var = np.sqrt(var) # to get standard deviation
plt = None # turn off plotting for now
if plt:
axis = np.arange(N)
plt.plot(axis, x[:, 0], 'k-', label='x_1')
plt.plot(axis, x[:, 1], 'k--', label='x_2')
plt.errorbar(axis, mean[:, 0], fmt='s',
label='mu_1') # yerror=var[:,0]
plt.errorbar(axis, mean[:, 1], fmt='D',
label='mu_2') # yerror=var[:,1]
plt.legend()
plt.show()
savemat(output_file, {
"Mu_py": mean.T,
"exec_time_pybayes": timer.spent[0]
},
oned_as='row')
def test_deepcopy(self):
"""Test that deep copying KF works as expected"""
o = pb.KalmanFilter(**self.setup_2) # original
c = deepcopy(o) # copy
self.assertEqual(type(c), type(o))
self.assertTrue(id(o) != id(c))
for (a, b) in [(o.A, c.A), (o.B, c.B), (o.C, c.C), (o.D, c.D), (o.Q, c.Q), (o.R, c.R)]:
self.assertArraysEqualNotSame(a, b)
# n, k, j do not need to be different as they are immutable
self.assertEqual(o.n, c.n)
self.assertEqual(o.k, c.k)
self.assertEqual(o.j, c.j)
self.assertTrue(id(o.P) != id(c.P))
self.assertArraysEqualNotSame(o.P.mu, c.P.mu) # this is better tested in
self.assertArraysEqualNotSame(o.P.R, c.P.R) # GaussPdf deepcopy test, but wont hurt here
self.assertTrue(id(o.S) != id(c.S))
def test_copy(self):
"""Test that copying KF works as expected"""
o = pb.KalmanFilter(**self.setup_1) # original
c = copy(o) # copy
self.assertEqual(type(c), type(o))
self.assertTrue(id(o) != id(c))
self.assertTrue(id(o.A) == id(c.A))
self.assertTrue(id(o.B) == id(c.B))
self.assertTrue(id(o.C) == id(c.C))
self.assertTrue(id(o.D) == id(c.D))
self.assertTrue(id(o.Q) == id(c.Q))
self.assertTrue(id(o.R) == id(c.R))
self.assertTrue(id(o.n) == id(c.n))
self.assertTrue(id(o.k) == id(c.k))
self.assertTrue(id(o.j) == id(c.j))
self.assertTrue(id(o.P) == id(c.P))
self.assertTrue(id(o.S) == id(c.S))
def test_copy(self):
"""Test that copying KF works as expected"""
o = pb.KalmanFilter(**self.setup_1) # original
c = copy(o) # copy
self.assertEqual(type(c), type(o))
self.assertNotEqual(id(o), id(c))
self.assertArraysSame(o.A, c.A)
self.assertArraysSame(o.B, c.B)
self.assertArraysSame(o.C, c.C)
self.assertArraysSame(o.D, c.D)
self.assertArraysSame(o.Q, c.Q)
self.assertArraysSame(o.R, c.R)
self.assertEqual(o.n, c.n)
self.assertEqual(o.k, c.k)
self.assertEqual(o.j, c.j)
self.assertEqual(id(o.P), id(c.P))
self.assertEqual(id(o.S), id(c.S))
def test_init(self):
k = pb.KalmanFilter(**self.setup_1)
self.assertEqual(type(k), pb.KalmanFilter)
l = pb.KalmanFilter(**self.setup_2)
self.assertEqual(type(l), pb.KalmanFilter)
try:
from scipy.io import loadmat, savemat
except ImportError, e:
raise StopIteration("Kalman filter stress needs scipy installed: " + str(e))
d = loadmat(input_file, struct_as_record=True, mat_dtype=True)
mu0 = np.reshape(d.pop('mu0'), (-1,)) # otherwise we would get 2D array of shape (1xN)
P0 = d.pop('P0')
y = d.pop('y').T
u = d.pop('u').T
#x = d.pop('x').T
x = None
gauss = pb.GaussPdf(mu0, P0)
kalman = pb.KalmanFilter(d['A'], d['B'], d['C'], d['D'], d['Q'], d['R'], gauss)
N = y.shape[0]
n = mu0.shape[0]
mean = np.zeros((N, n))
var = np.zeros((N, n))
timer.start()
for t in xrange(1, N): # the 1 start offset is intentional
kalman.bayes(y[t], u[t])
mean[t] = kalman.posterior().mean()
#var[t] = kalman.posterior().variance()
timer.stop()
var = np.sqrt(var) # to get standard deviation
plt = None # turn off plotting for now
