from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.1, 0.5, 0.1], [0.1, 0.1, 0.5], [0.5, 0.1, 0.1]]
alphabet = ['a', 'b', 'c']
B = [[0.8, 0.1, 0.1], [0.1, 0.8, 0.1], [0.1, 0.1, 0.8]]
my_markov_model = Markov_model(initial,
A,
B,
alphabet)
data = my_markov_model.generate_data(10)
my_proba_computer = Proba_computer(initial,
A,
B,
alphabet)
print [x['state'] for x in data]
print [x['obs'] for x in data]
forwards = my_proba_computer.compute_forward_probas([x['obs'] for x in data])
plt.subplot(311)
plt.imshow(forwards, cmap = cm.gray)
backwards = my_proba_computer.compute_backward_probas([x['obs'] for x in data])
plt.subplot(312)
plt.imshow(backwards, cmap = cm.gray)
import matplotlib.pyplot as plt
import matplotlib.cm as cm
from surrogate.Markov_model import Markov_model
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.3, 0.5, 0.3], [0.3, 0.3, 0.5], [0.5, 0.3, 0.3]]
alphabet = ['o', '*', 'p', 'h']
B = [[1.0, 0.5, 0.5, 0.5], [0.5, 1.0, 0.5, 0.5], [0.5, 0.5, 1.0, 0.5]]
my_markov_model = Markov_model(initial, A, B, alphabet)
data = my_markov_model.generate_data(20)
my_proba_computer = Proba_computer(initial, A, B, alphabet)
states = np.asarray([x['state'] for x in data])
observations = [x['obs'] for x in data]
gammas = my_proba_computer.compute_probas(observations)
epsilons = my_proba_computer.compute_epsilons(observations)
print epsilons
gammas_eps = np.zeros((epsilons.shape[0], epsilons.shape[2]))
for t in xrange(gammas_eps.shape[1]):
gammas_eps[:, t] = np.sum(epsilons[:, :, t], axis=1)
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.3, 0.5, 0.3], [0.3, 0.3, 0.5], [0.5, 0.3, 0.3]]
alphabet = ['o', '*', 'p', 'h']
B = [[1.0, 0.5, 0.5, 0.5], [0.5, 1.0, 0.5, 0.5], [0.5, 0.5, 1.0, 0.5]]
my_markov_model = Markov_model(initial,
A,
B,
alphabet)
datas = [my_markov_model.generate_data(100) for i in xrange(100)]
proba_cpter = Proba_computer(initial,
A,
B,
alphabet)
observations = [[x['obs'] for x in y] for y in datas]
initial = proba_cpter.initial
A = proba_cpter.A
B = proba_cpter.B
new_initial, new_A, new_B = proba_cpter.estimate_new_model_multi(observations)
print 'Initial'
print 'Model:'
print initial
print 'Estimated'
print new_initial
import matplotlib.pyplot as plt import matplotlib.cm as cm from surrogate.Markov_model import Markov_model from HMM_algos import Proba_computer initial = [0.1, 0.1, 0.1] A = [[0.1, 0.5, 0.1], [0.1, 0.1, 0.5], [0.5, 0.1, 0.1]] alphabet = ['a', 'b', 'c'] B = [[0.8, 0.1, 0.1], [0.1, 0.8, 0.1], [0.1, 0.1, 0.8]] my_markov_model = Markov_model(initial, A, B, alphabet) data = my_markov_model.generate_data(10) my_proba_computer = Proba_computer(initial, A, B, alphabet) print[x['state'] for x in data] print[x['obs'] for x in data] forwards = my_proba_computer.compute_forward_probas([x['obs'] for x in data]) plt.subplot(311) plt.imshow(forwards, cmap=cm.gray) backwards = my_proba_computer.compute_backward_probas([x['obs'] for x in data]) plt.subplot(312) plt.imshow(backwards, cmap=cm.gray) probas = my_proba_computer.compute_probas([x['obs'] for x in data])
from surrogate.Markov_model import Markov_model
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.1, 0.5, 0.1], [0.1, 0.1, 0.5], [0.5, 0.1, 0.1]]
alphabet = ['a', 'b', 'c']
B = [[1.0, 0.5, 0.5], [0.5, 1.0, 0.5], [0.5, 0.5, 1.0]]
my_markov_model = Markov_model(initial,
A,
B,
alphabet)
data = my_markov_model.generate_data(10)
my_forward_proba = Proba_computer(initial,
A,
B,
alphabet)
print [x['state'] for x in data]
print [x['obs'] for x in data]
forward_probas = my_forward_proba.compute_forward_probas([x['obs'] for x in data])
for i in xrange(forward_probas.shape[1]):
print forward_probas[:,i]
plt.imshow(forward_probas, cmap = cm.gray)
plt.clim()
plt.show()
import matplotlib.pyplot as plt
import matplotlib.cm as cm
from surrogate.Markov_model import Markov_model
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.3, 0.5, 0.3], [0.3, 0.3, 0.5], [0.5, 0.3, 0.3]]
alphabet = ["o", "*", "p", "h"]
B = [[1.0, 0.5, 0.5, 0.5], [0.5, 1.0, 0.5, 0.5], [0.5, 0.5, 1.0, 0.5]]
my_markov_model = Markov_model(initial, A, B, alphabet)
data = my_markov_model.generate_data(20)
my_proba_computer = Proba_computer(initial, A, B, alphabet)
states = np.asarray([x["state"] for x in data])
observations = [x["obs"] for x in data]
gammas = my_proba_computer.compute_probas(observations)
epsilons = my_proba_computer.compute_epsilons(observations)
print epsilons
gammas_eps = np.zeros((epsilons.shape[0], epsilons.shape[2]))
for t in xrange(gammas_eps.shape[1]):
gammas_eps[:, t] = np.sum(epsilons[:, :, t], axis=1)
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.3, 0.5, 0.3], [0.3, 0.3, 0.5], [0.5, 0.3, 0.3]]
alphabet = ['o', '*', 'p', 'h']
B = [[1.0, 0.5, 0.5, 0.5], [0.5, 1.0, 0.5, 0.5], [0.5, 0.5, 1.0, 0.5]]
my_markov_model = Markov_model(initial,
A,
B,
alphabet)
data = my_markov_model.generate_data(100)
proba_cpter = Proba_computer(initial,
A,
B,
alphabet)
states = np.asarray([x['state'] for x in data])
observations = [x['obs'] for x in data]
actual_initial = proba_cpter.initial
actual_A = proba_cpter.A
actual_B = proba_cpter.B
for i in xrange(100):
initial = proba_cpter.initial
A = proba_cpter.A
B = proba_cpter.B
new_initial, new_A, new_B = proba_cpter.estimate_new_model(observations)
print 'Initial'
from surrogate.Markov_model import Markov_model
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.1, 0.5, 0.1], [0.1, 0.1, 0.5], [0.5, 0.1, 0.1]]
alphabet = ['a', 'b', 'c']
B = [[1.0, 0.5, 0.5], [0.5, 1.0, 0.5], [0.5, 0.5, 1.0]]
my_markov_model = Markov_model(initial,
A,
B,
alphabet)
data = my_markov_model.generate_data(10)
my_forward_proba = Proba_computer(initial,
A,
B,
alphabet)
print [x['state'] for x in data]
print [x['obs'] for x in data]
forward_probas = my_forward_proba.compute_forward_probas([x['obs'] for x in data])
for i in xrange(forward_probas.shape[1]):
print forward_probas[:,i]
plt.imshow(forward_probas, cmap = cm.gray)
plt.clim()
plt.show()
import matplotlib.pyplot as plt import matplotlib.cm as cm from surrogate.Markov_model import Markov_model from HMM_algos import Proba_computer initial = [0.1, 0.1, 0.1] A = [[0.3, 0.5, 0.3], [0.3, 0.3, 0.5], [0.5, 0.3, 0.3]] alphabet = ['o', '*', 'p', 'h'] B = [[1.0, 0.5, 0.5, 0.5], [0.5, 1.0, 0.5, 0.5], [0.5, 0.5, 1.0, 0.5]] my_markov_model = Markov_model(initial, A, B, alphabet) datas = [my_markov_model.generate_data(100) for i in xrange(100)] proba_cpter = Proba_computer(initial, A, B, alphabet) observations = [[x['obs'] for x in y] for y in datas] initial = proba_cpter.initial A = proba_cpter.A B = proba_cpter.B new_initial, new_A, new_B = proba_cpter.estimate_new_model_multi(observations) print 'Initial' print 'Model:' print initial print 'Estimated' print new_initial print '\n'
import matplotlib.pyplot as plt
import matplotlib.cm as cm
from surrogate.Markov_model import Markov_model
from HMM_algos import Proba_computer
initial = [0.1, 0.1, 0.1]
A = [[0.3, 0.5, 0.3], [0.3, 0.3, 0.5], [0.5, 0.3, 0.3]]
alphabet = ["o", "*", "p", "h"]
B = [[1.0, 0.5, 0.5, 0.5], [0.5, 1.0, 0.5, 0.5], [0.5, 0.5, 1.0, 0.5]]
my_markov_model = Markov_model(initial, A, B, alphabet)
data = my_markov_model.generate_data(100)
proba_cpter = Proba_computer(initial, A, B, alphabet)
states = np.asarray([x["state"] for x in data])
observations = [x["obs"] for x in data]
actual_initial = proba_cpter.initial
actual_A = proba_cpter.A
actual_B = proba_cpter.B
for i in xrange(100):
initial = proba_cpter.initial
A = proba_cpter.A
B = proba_cpter.B
new_initial, new_A, new_B = proba_cpter.estimate_new_model(observations)
print "Initial"
print "Model:"