def test_init(self):
"""
Test the creation of Bernoulli nodes.
"""
# Some simple initializations
X = Bernoulli(0.5)
X = Bernoulli(Beta([2, 3]))
# Check that plates are correct
X = Bernoulli(0.7, plates=(4, 3))
self.assertEqual(X.plates, (4, 3))
X = Bernoulli(0.7 * np.ones((4, 3)))
self.assertEqual(X.plates, (4, 3))
X = Bernoulli(Beta([4, 3], plates=(4, 3)))
self.assertEqual(X.plates, (4, 3))
# Invalid probability
self.assertRaises(ValueError, Bernoulli, -0.5)
self.assertRaises(ValueError, Bernoulli, 1.5)
# Inconsistent plates
self.assertRaises(ValueError,
Bernoulli,
0.5 * np.ones(4),
plates=(3, ))
# Explicit plates too small
self.assertRaises(ValueError,
Bernoulli,
0.5 * np.ones(4),
plates=(1, ))
pass
def _setup_bernoulli_mixture():
"""
Setup code for the hinton tests.
This code is from http://www.bayespy.org/examples/bmm.html
"""
np.random.seed(1)
p0 = [0.1, 0.9, 0.1, 0.9, 0.1, 0.9, 0.1, 0.9, 0.1, 0.9]
p1 = [0.1, 0.1, 0.1, 0.1, 0.1, 0.9, 0.9, 0.9, 0.9, 0.9]
p2 = [0.9, 0.9, 0.9, 0.9, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1]
p = np.array([p0, p1, p2])
z = random.categorical([1 / 3, 1 / 3, 1 / 3], size=100)
x = random.bernoulli(p[z])
N = 100
D = 10
K = 10
R = Dirichlet(K * [1e-5], name='R')
Z = Categorical(R, plates=(N, 1), name='Z')
P = Beta([0.5, 0.5], plates=(D, K), name='P')
X = Mixture(Z, Bernoulli, P)
Q = VB(Z, R, X, P)
P.initialize_from_random()
X.observe(x)
Q.update(repeat=1000)
return (R, P, Z)
def test_init(self):
"""
Test the creation of beta nodes.
"""
# Some simple initializations
p = Beta([1.5, 4.2])
# Check that plates are correct
p = Beta([2, 3], plates=(4, 3))
self.assertEqual(p.plates, (4, 3))
p = Beta(np.ones((4, 3, 2)))
self.assertEqual(p.plates, (4, 3))
# Parent not a vector
self.assertRaises(ValueError, Beta, 4)
# Parent vector has wrong shape
self.assertRaises(ValueError, Beta, [4])
self.assertRaises(ValueError, Beta, [4, 4, 4])
# Parent vector has invalid values
self.assertRaises(ValueError, Beta, [-2, 3])
# Plates inconsistent
self.assertRaises(ValueError, Beta, np.ones((4, 2)), plates=(3, ))
# Explicit plates too small
self.assertRaises(ValueError, Beta, np.ones((4, 2)), plates=(1, ))
pass
def test_moments(self):
"""
Test the moments of Bernoulli nodes.
"""
# Simple test
X = Bernoulli(0.7)
u = X._message_to_child()
self.assertEqual(len(u), 1)
self.assertAllClose(u[0], 0.7)
# Test plates in p
p = np.random.rand(3)
X = Bernoulli(p)
u = X._message_to_child()
self.assertAllClose(u[0], p)
# Test with beta prior
P = Beta([7, 3])
logp = P._message_to_child()[0]
p0 = np.exp(logp[0]) / (np.exp(logp[0]) + np.exp(logp[1]))
X = Bernoulli(P)
u = X._message_to_child()
self.assertAllClose(u[0], p0)
# Test with broadcasted plates
P = Beta([7, 3], plates=(10, ))
X = Bernoulli(P)
u = X._message_to_child()
self.assertAllClose(u[0] * np.ones(X.get_shape(0)), p0 * np.ones(10))
pass
def test_moments(self):
"""
Test the moments of Bernoulli nodes.
"""
# Simple test
X = Bernoulli(0.7)
u = X._message_to_child()
self.assertEqual(len(u), 1)
self.assertAllClose(u[0], 0.7)
# Test plates in p
p = np.random.rand(3)
X = Bernoulli(p)
u = X._message_to_child()
self.assertAllClose(u[0], p)
# Test with beta prior
P = Beta([7, 3])
logp = P._message_to_child()[0]
p0 = np.exp(logp[0]) / (np.exp(logp[0]) + np.exp(logp[1]))
X = Bernoulli(P)
u = X._message_to_child()
self.assertAllClose(u[0], p0)
# Test with broadcasted plates
P = Beta([7, 3], plates=(10,))
X = Bernoulli(P)
u = X._message_to_child()
self.assertAllClose(u[0] * np.ones(X.get_shape(0)), p0 * np.ones(10))
pass
def test_moments(self):
"""
Test the moments of binomial nodes.
"""
# Simple test
X = Binomial(1, 0.7)
u = X._message_to_child()
self.assertEqual(len(u), 1)
self.assertAllClose(u[0],
0.7)
# Test n
X = Binomial(10, 0.7)
u = X._message_to_child()
self.assertAllClose(u[0],
10*0.7)
# Test plates in p
n = np.random.randint(1, 10)
p = np.random.rand(3)
X = Binomial(n, p)
u = X._message_to_child()
self.assertAllClose(u[0],
p*n)
# Test plates in n
n = np.random.randint(1, 10, size=(3,))
p = np.random.rand()
X = Binomial(n, p)
u = X._message_to_child()
self.assertAllClose(u[0],
p*n)
# Test plates in p and n
n = np.random.randint(1, 10, size=(4,1))
p = np.random.rand(3)
X = Binomial(n, p)
u = X._message_to_child()
self.assertAllClose(u[0],
p*n)
# Test with beta prior
P = Beta([7, 3])
logp = P._message_to_child()[0]
p0 = np.exp(logp[0]) / (np.exp(logp[0]) + np.exp(logp[1]))
X = Binomial(1, P)
u = X._message_to_child()
self.assertAllClose(u[0],
p0)
# Test with broadcasted plates
P = Beta([7, 3], plates=(10,))
X = Binomial(5, P)
u = X._message_to_child()
self.assertAllClose(u[0] * np.ones(X.get_shape(0)),
5*p0*np.ones(10))
pass
def test_moments(self):
"""
Test the moments of beta nodes.
"""
p = Beta([2, 3])
u = p._message_to_child()
self.assertAllClose(u[0], special.psi([2, 3]) - special.psi(2 + 3))
pass
def test_moments(self):
"""
Test the moments of beta nodes.
"""
p = Beta([2, 3])
u = p._message_to_child()
self.assertAllClose(u[0], special.psi([2, 3]) - special.psi(2 + 3))
pass
def test_moments(self):
"""
Test the moments of binomial nodes.
"""
# Simple test
X = Binomial(1, 0.7)
u = X._message_to_child()
self.assertEqual(len(u), 1)
self.assertAllClose(u[0], 0.7)
# Test n
X = Binomial(10, 0.7)
u = X._message_to_child()
self.assertAllClose(u[0], 10 * 0.7)
# Test plates in p
n = np.random.randint(1, 10)
p = np.random.rand(3)
X = Binomial(n, p)
u = X._message_to_child()
self.assertAllClose(u[0], p * n)
# Test plates in n
n = np.random.randint(1, 10, size=(3, ))
p = np.random.rand()
X = Binomial(n, p)
u = X._message_to_child()
self.assertAllClose(u[0], p * n)
# Test plates in p and n
n = np.random.randint(1, 10, size=(4, 1))
p = np.random.rand(3)
X = Binomial(n, p)
u = X._message_to_child()
self.assertAllClose(u[0], p * n)
# Test with beta prior
P = Beta([7, 3])
logp = P._message_to_child()[0]
p0 = np.exp(logp[0]) / (np.exp(logp[0]) + np.exp(logp[1]))
X = Binomial(1, P)
u = X._message_to_child()
self.assertAllClose(u[0], p0)
# Test with broadcasted plates
P = Beta([7, 3], plates=(10, ))
X = Binomial(5, P)
u = X._message_to_child()
self.assertAllClose(u[0] * np.ones(X.get_shape(0)),
5 * p0 * np.ones(10))
pass
def _run(self, x, K=25, beta=0.5, alpha=0.00001, hinton_plot=False, end=False):
'''Only to be used when doing parameter optimization.'''
self.participant_list = x[0]
N = len(x[0]) #number of data points (i.e. WCS participants)
D = np.shape(x[1])[1] #number of features
#K = 20 #number of initial clusters
R = Dirichlet(K*[alpha],
name='R')
Z = Categorical(R,
plates=(N,1),
name='Z')
P = Beta([beta, beta],
plates=(D,K),
name='P')
X = Mixture(Z, Bernoulli, P)
Q = VB(Z, R, X, P)
P.initialize_from_random()
X.observe(x[1])
Q.update(repeat=1000)
log_likelihood = Q.L[Q.iter-1]
if hinton_plot:
bpplt.hinton(Z)
bpplt.pyplot.show()
bpplt.hinton(R)
bpplt.pyplot.show()
#Get the weight matrix stored in Z (weights determine which cluster data point belongs to)
z = Z._message_to_child()[0]
z = z * np.ones(Z.plates+(1,))
z = np.squeeze(z)
self.z = z
#Get the weights stored in R (proportional to the size of the clusters)
r = np.exp(R._message_to_child()[0])
r = r * np.ones(R.plates+(1,))
r = np.squeeze(r)
self.r = r
#Get the cluster assignment of each data point
self.c_assign = np.argmax(self.z, axis=1)
return log_likelihood
def test_init(self):
"""
Test the creation of binomial nodes.
"""
# Some simple initializations
X = Binomial(10, 0.5)
X = Binomial(10, Beta([2, 3]))
# Check that plates are correct
X = Binomial(10, 0.7, plates=(4, 3))
self.assertEqual(X.plates, (4, 3))
X = Binomial(10, 0.7 * np.ones((4, 3)))
self.assertEqual(X.plates, (4, 3))
n = np.ones((4, 3), dtype=np.int)
X = Binomial(n, 0.7)
self.assertEqual(X.plates, (4, 3))
X = Binomial(10, Beta([4, 3], plates=(4, 3)))
self.assertEqual(X.plates, (4, 3))
# Invalid probability
self.assertRaises(ValueError, Binomial, 10, -0.5)
self.assertRaises(ValueError, Binomial, 10, 1.5)
# Invalid number of trials
self.assertRaises(ValueError, Binomial, -1, 0.5)
self.assertRaises(ValueError, Binomial, 8.5, 0.5)
# Inconsistent plates
self.assertRaises(ValueError,
Binomial,
10,
0.5 * np.ones(4),
plates=(3, ))
# Explicit plates too small
self.assertRaises(ValueError,
Binomial,
10,
0.5 * np.ones(4),
plates=(1, ))
pass
def test_random(self):
"""
Test random sampling of beta nodes.
"""
p = Beta([1e20, 3e20])
x = p.random()
self.assertAllClose(x, 0.25)
p = Beta([[1e20, 3e20], [1e20, 1e20]])
x = p.random()
self.assertAllClose(x, [0.25, 0.5])
p = Beta([1e20, 3e20], plates=(3, ))
x = p.random()
self.assertAllClose(x, [0.25, 0.25, 0.25])
pass
def _setup_bernoulli_mixture():
"""
Setup code for the hinton tests.
This code is from http://www.bayespy.org/examples/bmm.html
"""
np.random.seed(1)
p0 = [0.1, 0.9, 0.1, 0.9, 0.1, 0.9, 0.1, 0.9, 0.1, 0.9]
p1 = [0.1, 0.1, 0.1, 0.1, 0.1, 0.9, 0.9, 0.9, 0.9, 0.9]
p2 = [0.9, 0.9, 0.9, 0.9, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1]
p = np.array([p0, p1, p2])
z = random.categorical([1/3, 1/3, 1/3], size=100)
x = random.bernoulli(p[z])
N = 100
D = 10
K = 10
R = Dirichlet(K*[1e-5],
name='R')
Z = Categorical(R,
plates=(N,1),
name='Z')
P = Beta([0.5, 0.5],
plates=(D,K),
name='P')
X = Mixture(Z, Bernoulli, P)
Q = VB(Z, R, X, P)
P.initialize_from_random()
X.observe(x)
Q.update(repeat=1000)
return (R,P,Z)
def test_random(self):
"""
Test random sampling of beta nodes.
"""
p = Beta([1e20, 3e20])
x = p.random()
self.assertAllClose(x, 0.25)
p = Beta([[1e20, 3e20], [1e20, 1e20]])
x = p.random()
self.assertAllClose(x, [0.25, 0.5])
p = Beta([1e20, 3e20], plates=(3,))
x = p.random()
self.assertAllClose(x, [0.25, 0.25, 0.25])
pass
import numpy numpy.random.seed(1) p0 = [0.1, 0.9, 0.1, 0.9, 0.1, 0.9, 0.1, 0.9, 0.1, 0.9] p1 = [0.1, 0.1, 0.1, 0.1, 0.1, 0.9, 0.9, 0.9, 0.9, 0.9] p2 = [0.9, 0.9, 0.9, 0.9, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1] import numpy as np p = np.array([p0, p1, p2]) from bayespy.utils import random z = random.categorical([1 / 3, 1 / 3, 1 / 3], size=100) x = random.bernoulli(p[z]) N = 100 D = 10 K = 10 from bayespy.nodes import Categorical, Dirichlet R = Dirichlet(K * [1e-5], name='R') Z = Categorical(R, plates=(N, 1), name='Z') from bayespy.nodes import Beta P = Beta([0.5, 0.5], plates=(D, K), name='P') from bayespy.nodes import Mixture, Bernoulli X = Mixture(Z, Bernoulli, P) from bayespy.inference import VB Q = VB(Z, R, X, P) P.initialize_from_random() X.observe(x) Q.update(repeat=1000) import bayespy.plot as bpplt bpplt.hinton(P) bpplt.pyplot.show()
p0 = [0.1, 0.1, 0.1, 0.9, 0.9, 0.9, 0.1, 0.1, 0.1, 0.1]
p1 = [0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.9, 0.9, 0.9, 0.9]
p2 = [0.9, 0.9, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1]
p = np.array([p0, p1, p2])
z = random.categorical([1/3, 1/3, 1/3], size=100)
x = random.bernoulli(p[z])
N = 100
D = 10
K = 3
R = Dirichlet(K*[1e-5],name='R')
Z = Categorical(R,plates=(N,1),name='Z')
P = Beta([0.5, 0.5],plates=(D,K),name='P')
X = Mixture(Z, Bernoulli, P)
Q = VB(Z, R, X, P)
P.initialize_from_random()
X.observe(x)
Q.update(repeat=1000)
#print(" P:")
#print( P.get_moments() )
#print(" R:")
#print( R.get_moments() )
print(" Z:")
def run(self, K=25, beta=0.5, alpha=0.00001, foci_thresh=0, num_neigh=4, hinton_plot=False, end=False):
'''Performs one run of the BBDP according to the specified parameters.'''
print("Transforming WCS participant data into binary vectors...")
x = u.transform_data_all(self.langs, norm=False, end=end, foci=True, foci_thresh=foci_thresh, num_neigh=num_neigh)
print("Finished transforming participant data")
self.participant_list = x[0]
N = len(x[0]) #number of data points (i.e. WCS participants)
D = np.shape(x[1])[1] #number of features
#K = 20 #number of initial clusters
R = Dirichlet(K*[alpha],
name='R')
Z = Categorical(R,
plates=(N,1),
name='Z')
P = Beta([beta, beta],
plates=(D,K),
name='P')
X = Mixture(Z, Bernoulli, P)
Q = VB(Z, R, X, P)
P.initialize_from_random()
X.observe(x[1])
Q.update(repeat=1000)
if hinton_plot:
bpplt.hinton(Z)
bpplt.pyplot.show()
bpplt.hinton(R)
bpplt.pyplot.show()
#Get the weight matrix stored in Z (weights determine which cluster data point belongs to)
z = Z._message_to_child()[0]
z = z * np.ones(Z.plates+(1,))
z = np.squeeze(z)
self.z = z
#Get the weights stored in R (proportional to the size of the clusters)
r = np.exp(R._message_to_child()[0])
r = r * np.ones(R.plates+(1,))
r = np.squeeze(r)
self.r = r
#Get the cluster assignment of each data point
self.c_assign = np.argmax(self.z, axis=1)
#Write cluster results to a file
if self.write_to_file:
if end:
save_path = "cluster_results_end_K={}_B={}_a={}_t={}_nn={}".format(K, beta, alpha, foci_thresh, num_neigh)
else:
save_path = "cluster_results_K={}_B={}_a={}_t={}_nn={}".format(K, beta, alpha, foci_thresh, num_neigh)
while path.exists(save_path+".txt"):
#save_path already exists
try:
old_file_num = int(save_path[save_path.find('(')+1:-1])
new_file_num = old_file_num + 1
save_path = save_path[0:save_path.find('(')] + '(' + str(new_file_num) + ')'
except ValueError:
save_path = save_path + " (1)"
self.save_path = save_path
file = open(path.abspath(self.save_path+".txt"), 'w')
#Write cluster assignment matrix Z (gives the probability that observation i belongs to cluster j)
if 'Z' not in self.in_file:
for i in range(len(self.z)):
line = "\t".join([str(x) for x in self.z[i]]) + "\n"
file.write(line)
file.write('---Z\n')
self.in_file.append('Z')
#Write cluster weights matrix R (proportional to the size of the resulting clusters)
if 'R' not in self.in_file:
line = "\t".join([str(x) for x in self.r]) + "\n"
file.write(line)
file.write('---R\n')
self.in_file.append('R')
#Write deterministic cluster assignments with the corresponding participant key
if 'C' not in self.in_file:
line1 = "\t".join([str(x) for x in self.participant_list]) + "\n"
line2 = "\t".join([str(x) for x in self.c_assign]) + "\n"
file.write(line1)
file.write(line2)
file.write('---C\n')
self.in_file.append('C')
file.close()
return self.c_assign