def tets_distributions(): # not distributions d = numpy.ones((3,3)) assert(mkm.is_distribution(d) == False) d = mkm.delta_distribution(10) d[1] = 1 assert(mkm.is_distribution(d) == False) assert(mkm.is_distribution(mkm.delta_distribution(10),n=11) == False) assert(mkm.is_distribution(mkm.delta_distribution(10),n=10) == True) assert(mkm.relative_error(numpy.array([0.0001,1,1e-16]),numpy.array([0.001,1,1e-16]))>8.0)
def add_distributions(self,d):
# vector shape
if d.ndim == 1:
if mkm.is_distribution(d) == False:
raise Exception('not a probability distribution')
new = {0: d}
self.distributions.append(new)
# matrix shape
else:
for i in range(d.shape[0]):
if mkm.is_distribution(d[i,:]) == False:
raise Exception('not a probability distribution')
new = {0: d[i,:]}
self.distributions.append(new)
def set_stationary(self, d):
""" Set the stationary distribution.
d: the stationary distribution
"""
if mkm.is_distribution(d) == False:
raise Exception('not a probability distribution')
self.sd = d
def set_stationary(self,d):
""" Set the stationary distribution.
d: the stationary distribution
"""
if mkm.is_distribution(d) == False:
raise Exception('not a probability distribution')
self.sd = d
def add_iteration(self, index, t, iteration):
""" Add an iteration for a distribution.
In the terminology of this class the iteration for a distribution
at time t is defined by distribution*P^{t}.
index: index of the distribution
t: iteration time
iteration: the iteration
"""
if mkm.is_distribution(iteration) == False:
raise Exception('not a probability distribution')
self.distributions[index][t] = iteration
def add_distributions(self, d):
""" Add a number of distributions to the Markov chain.
A distribution in the sense of this class is an initial
distribution on the state space.
d: An ndarray where each row is a distribution
"""
# vector shape
if d.ndim == 1:
if mkm.is_distribution(d) == False:
raise Exception('not a probability distribution')
new = {0: d}
self.distributions.append(new)
# matrix shape
else:
for i in range(d.shape[0]):
if mkm.is_distribution(d[i, :]) == False:
raise Exception('not a probability distribution')
new = {0: d[i, :]}
self.distributions.append(new)
def add_iteration(self,index,t,iteration):
""" Add an iteration for a distribution.
In the terminology of this class the iteration for a distribution
at time t is defined by distribution*P^{t}.
index: index of the distribution
t: iteration time
iteration: the iteration
"""
if mkm.is_distribution(iteration) == False:
raise Exception('not a probability distribution')
self.distributions[index][t] = iteration
def add_distributions(self,d):
""" Add a number of distributions to the Markov chain.
A distribution in the sense of this class is an initial
distribution on the state space.
d: An ndarray where each row is a distribution
"""
# vector shape
if d.ndim == 1:
if mkm.is_distribution(d) == False:
raise Exception('not a probability distribution')
new = {0: d}
self.distributions.append(new)
# matrix shape
else:
for i in range(d.shape[0]):
if mkm.is_distribution(d[i,:]) == False:
raise Exception('not a probability distribution')
new = {0: d[i,:]}
self.distributions.append(new)
def add_iteration(self,index,t,iteration):
if mkm.is_distribution(iteration) == False:
raise Exception('not a probability distribution')
self.distributions[index][t] = iteration
def set_stationary_distribution(self,d):
if mkm.is_distribution(d) == False:
raise Exception('not a probability distribution')
self.sd = d