|
|
|
|
import numpy as np
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class ConditionalIntensityMatrix:
|
|
|
|
|
|
|
|
|
|
def __init__(self, dimension, state_residence_times, state_transition_matrix):
|
|
|
|
|
self.state_residence_times = state_residence_times
|
|
|
|
|
self.state_transition_matrix = state_transition_matrix
|
|
|
|
|
#self.cim = np.zeros(shape=(dimension, dimension), dtype=float)
|
|
|
|
|
self.cim = self.state_transition_matrix.astype(np.float)
|
|
|
|
|
|
|
|
|
|
def update_state_transition_count(self, element_indx):
|
|
|
|
|
#print(element_indx)
|
|
|
|
|
#self.state_transition_matrix[element_indx[0]][element_indx[1]] += 1
|
|
|
|
|
self.state_transition_matrix[element_indx] += 1
|
|
|
|
|
|
|
|
|
|
def update_state_residence_time_for_state(self, state, time):
|
|
|
|
|
#print("Time updating In state", state, time)
|
|
|
|
|
self.state_residence_times[state] += time
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def compute_cim_coefficients(self):
|
|
|
|
|
np.fill_diagonal(self.cim, self.cim.diagonal() * -1)
|
|
|
|
|
self.cim = ((self.cim.T + 1) / (self.state_residence_times + 1)).T
|
|
|
|
|
#np.fill_diagonal(self.state_transition_matrix, 0)
|
|
|
|
|
|
|
|
|
|
def __repr__(self):
|
|
|
|
|
return 'CIM:\n' + str(self.cim)
|
|
|
|
|
|
