def do_gibbs_best(self, value):
"""Run Gibbs sampling to produce the best labelling for
given input value"""
xs = parse_line(value)
ys = computeGibbsBestSequence(self.crf, getCRFBlocks,
chooseGibbsCRF, xs, 10000)
print pretty_format(ys)
def do_gibbs_dist(self, value):
"""Run Gibbs sampling to produce the best labelling for
given input value"""
xs = parse_line(value)
ys = computeGibbsProbabilities(self.crf, getCRFBlocks,
chooseGibbsCRF, xs, 10000)
for label, pr in ys.most_common(10):
print pr, '\t', pretty_format(label)
def do_gibbs_best(self, value):
"""Run Gibbs sampling to produce the best labelling for
given input value"""
xs = parse_line(value)
ys = computeGibbsBestSequence(
self.crf,
getCRFBlocks,
chooseGibbsCRF,
xs,
10000 )
print pretty_format(ys)
def do_gibbs_dist(self, value):
"""Run Gibbs sampling to produce the best labelling for
given input value"""
xs = parse_line(value)
ys = computeGibbsProbabilities(
self.crf,
getCRFBlocks,
chooseGibbsCRF,
xs,
10000 )
for label, pr in ys.most_common(10):
print pr, '\t', pretty_format(label)
def do_viterbi(self, value):
"""Run the viterbi algorithm on input to produce the most
likely labelling"""
xs = parse_line(value)
ys = computeViterbi(crf, xs)
print pretty_format(ys)
def do_viterbi(self, value):
"""Run the viterbi algorithm on input to produce the most
likely labelling"""
xs = parse_line(value)
ys = computeViterbi(crf, xs)
print pretty_format(ys)