def _sample_forwards_log(trans_matrix_docnum, log_likelihoods):
A = trans_matrix_docnum
aBl = log_likelihoods
T = aBl.shape[0]
stateseq = np.empty(T,dtype=np.int32)
nextstate_unsmoothed = trans_matrix_docnum
for idx in xrange(T):
logdomain = aBl[idx]
logdomain[nextstate_unsmoothed == 0] = -np.inf
if np.any(np.isfinite(logdomain)):
stateseq[idx] = sample_discrete(nextstate_unsmoothed * np.exp(logdomain - np.amax(logdomain)))
else:
stateseq[idx] = sample_discrete(nextstate_unsmoothed)
nextstate_unsmoothed = trans_matrix_docnum
return stateseq
def _sample_forwards_log(trans_matrix_docnum, log_likelihoods):
A = trans_matrix_docnum
aBl = log_likelihoods
T = aBl.shape[0]
stateseq = np.empty(T, dtype=np.int32)
nextstate_unsmoothed = trans_matrix_docnum
for idx in xrange(T):
logdomain = aBl[idx]
logdomain[nextstate_unsmoothed == 0] = -np.inf
if np.any(np.isfinite(logdomain)):
stateseq[idx] = sample_discrete(
nextstate_unsmoothed *
np.exp(logdomain - np.amax(logdomain)))
else:
stateseq[idx] = sample_discrete(nextstate_unsmoothed)
nextstate_unsmoothed = trans_matrix_docnum
return stateseq
def generate_states(self):
T = self.T
doc_num = self.doc_num
state_distn = self.trans_matrix[doc_num, :]
stateseq = np.zeros(T, dtype=np.int32)
for idx in xrange(T):
stateseq[idx] = sample_discrete(state_distn)
self.stateseq = stateseq
return stateseq
def generate_states(self):
T = self.T
doc_num = self.doc_num
state_distn = self.trans_matrix[doc_num, :]
stateseq = np.zeros(T,dtype=np.int32)
for idx in xrange(T):
stateseq[idx] = sample_discrete(state_distn)
self.stateseq = stateseq
return stateseq
def rvs_given_less_than(self,x,num):
pmf = self.pmf(np.arange(1,x))
return sample_discrete(pmf,num)+1
def rvs_given_less_than(self,x,num):
pmf = self.pmf(np.arange(1,x))
return sample_discrete(pmf,num)+1