def viterbi(HMM, ev, prior):
"""[Equation 15.11]
Viterbi algorithm to find the most likely sequence. Computes the best path,
given an HMM model and a sequence of observations."""
t = len(ev)
ev.insert(0, None)
m = [[0.0, 0.0] for _ in range(len(ev) - 1)]
# the recursion is initialized with m1 = forward(P(X0), e1)
m[0] = forward(HMM, prior, ev[1])
for i in range(1, t):
m[i] = element_wise_product(HMM.sensor_dist(ev[i + 1]), [
max(element_wise_product(HMM.transition_model[0], m[i - 1])),
max(element_wise_product(HMM.transition_model[1], m[i - 1]))
])
path = [0.0] * (len(ev) - 1)
# the construction of the most likely sequence starts in the final state with the largest probability,
# and runs backwards; the algorithm needs to store for each xt its best predecessor xt-1
for i in range(t, -1, -1):
path[i - 1] = max(m[i - 1])
return path
def forward_backward(HMM, ev):
"""
[Figure 15.4]
Forward-Backward algorithm for smoothing. Computes posterior probabilities
of a sequence of states given a sequence of observations."""
t = len(ev)
ev.insert(0, None) # to make the code look similar to pseudo code
fv = [[0.0, 0.0] for _ in range(len(ev))]
b = [1.0, 1.0]
bv = [b] # we don't need bv; but we will have a list of all backward messages here
sv = [[0, 0] for _ in range(len(ev))]
fv[0] = HMM.prior
for i in range(1, t + 1):
fv[i] = forward(HMM, fv[i - 1], ev[i])
for i in range(t, -1, -1):
sv[i - 1] = normalize(element_wise_product(fv[i], b))
b = backward(HMM, b, ev[i])
bv.append(b)
sv = sv[::-1]
return sv
def forward(HMM, fv, ev):
prediction = vector_add(
scalar_vector_product(fv[0], HMM.transition_model[0]),
scalar_vector_product(fv[1], HMM.transition_model[1]))
sensor_dist = HMM.sensor_dist(ev)
return normalize(element_wise_product(sensor_dist, prediction))
def backward(HMM, b, ev):
sensor_dist = HMM.sensor_dist(ev)
prediction = element_wise_product(sensor_dist, b)
return normalize(
vector_add(
scalar_vector_product(prediction[0], HMM.transition_model[0]),
scalar_vector_product(prediction[1], HMM.transition_model[1])))
def viterbi(HMM, ev):
"""
[Equation 15.11]
Viterbi algorithm to find the most likely sequence. Computes the best path and the
corresponding probabilities, given an HMM model and a sequence of observations.
"""
t = len(ev)
ev = ev.copy()
ev.insert(0, None)
m = [[0.0, 0.0] for _ in range(len(ev) - 1)]
# the recursion is initialized with m1 = forward(P(X0), e1)
m[0] = forward(HMM, HMM.prior, ev[1])
# keep track of maximizing predecessors
backtracking_graph = []
for i in range(1, t):
m[i] = element_wise_product(HMM.sensor_dist(ev[i + 1]), [
max(element_wise_product(HMM.transition_model[0], m[i - 1])),
max(element_wise_product(HMM.transition_model[1], m[i - 1]))
])
backtracking_graph.append([
np.argmax(element_wise_product(HMM.transition_model[0], m[i - 1])),
np.argmax(element_wise_product(HMM.transition_model[1], m[i - 1]))
])
# computed probabilities
ml_probabilities = [0.0] * (len(ev) - 1)
# most likely sequence
ml_path = [True] * (len(ev) - 1)
# the construction of the most likely sequence starts in the final state with the largest probability, and
# runs backwards; the algorithm needs to store for each xt its predecessor xt-1 maximizing its probability
i_max = np.argmax(m[-1])
for i in range(t - 1, -1, -1):
ml_probabilities[i] = m[i][i_max]
ml_path[i] = True if i_max == 0 else False
if i > 0:
i_max = backtracking_graph[i - 1][i_max]
return ml_path, ml_probabilities
def forward_backward(HMM, ev, prior):
"""Algoritmo forward-backward para suavização. Calcula probabilidades posteriores
De uma seqüência de estados dada uma seqüência de observações."""
t = len(ev)
ev.insert(0, None)
fv = [[0.0, 0.0] for i in range(len(ev))]
b = [1.0, 1.0]
bv = [b]
sv = [[0, 0] for i in range(len(ev))]
fv[0] = prior
for i in range(1, t + 1):
fv[i] = forward(HMM, fv[i - 1], ev[i])
for i in range(t, -1, -1):
sv[i - 1] = normalize(element_wise_product(fv[i], b))
b = backward(HMM, b, ev[i])
bv.append(b)
sv = sv[::-1]
return sv
def forward_backward(HMM, ev, prior):
"""[Figure 15.4]
Forward-Backward algorithm for smoothing. Computes posterior probabilities
of a sequence of states given a sequence of observations."""
t = len(ev)
ev.insert(0, None) # to make the code look similar to pseudo code
fv = [[0.0, 0.0] for i in range(len(ev))]
b = [1.0, 1.0]
bv = [b] # we don't need bv; but we will have a list of all backward messages here
sv = [[0, 0] for i in range(len(ev))]
fv[0] = prior
for i in range(1, t + 1):
fv[i] = forward(HMM, fv[i - 1], ev[i])
for i in range(t, -1, -1):
sv[i - 1] = normalize(element_wise_product(fv[i], b))
b = backward(HMM, b, ev[i])
bv.append(b)
sv = sv[::-1]
return(sv)
def backward(HMM, b, ev):
sensor_dist = HMM.sensor_dist(ev)
prediction = element_wise_product(sensor_dist, b)
return(normalize(vector_add(scalar_vector_product(prediction[0], HMM.transition_model[0]),
scalar_vector_product(prediction[1], HMM.transition_model[1]))))
def forward(HMM, fv, ev):
prediction = vector_add(scalar_vector_product(fv[0], HMM.transition_model[0]),
scalar_vector_product(fv[1], HMM.transition_model[1]))
sensor_dist = HMM.sensor_dist(ev)
return(normalize(element_wise_product(sensor_dist, prediction)))
