def run(self, data, regression=None, resources=None):
if data.ndim < 2:
raise StandardError, "Argument 'data' must be a 2D numpy array."
scale = resources.get("scale_for_data_standardization", 2)
for i in range(data.shape[1]):
data[:,i] = (data[:,i] - data[:,i].mean())/(scale*ndimage.standard_deviation(data[:,i]))
logger.log_status('Data has been standardized using (X-mean(X))/(%s*sd(X))' % scale)
return estimate_linear_regression.run(self, data, regression=regression, resources=resources)
def run(self, data, regression=None, resources=None):
if data.ndim < 2:
raise StandardError, "Argument 'data' must be a 2D numpy array."
scale = resources.get("scale_for_data_standardization", 2)
for i in range(data.shape[1]):
data[:, i] = (data[:, i] - data[:, i].mean()) / (
scale * ndimage.standard_deviation(data[:, i]))
logger.log_status(
'Data has been standardized using (X-mean(X))/(%s*sd(X))' % scale)
return estimate_linear_regression.run(self,
data,
regression=regression,
resources=resources)
def run(self, data, upc_sequence, resources=None):
CLOSE = 0.01
self.mnl_probabilities=upc_sequence.probability_class
self.bhhh_estimation = bhhh_mnl_estimation()
modified_upc_sequence = UPCFactory().get_model(
utilities=None, probabilities="opus_core.mnl_probabilities", choices=None)
modified_upc_sequence.utility_class = upc_sequence.utility_class
result = self.bhhh_estimation.run(data, modified_upc_sequence, resources)
probability = modified_upc_sequence.get_probabilities()
probability_0 = probability
resources.check_obligatory_keys(["capacity"])
supply = resources["capacity"]
if not isinstance(supply, ndarray):
supply = array(supply)
nsupply = supply.size
max_iter = resources.get("max_iterations", None)
if max_iter == None:
max_iter = 100 # default
index = resources.get("index", None)
if index == None:
index = arange(nsupply)
neqs = probability.shape[1]
nobs = probability.shape[0]
if index.ndim <= 1:
index = repeat(reshape(index, (1,index.shape[0])), nobs)
resources.merge({"index":index})
# WARNING: THE SCALING OF DEMAND IS HARD CODED AND NEEDS TO BE MADE AN ARGUMENT
# scale demand to represent 100% from a 0.2% sample
demand = self.mnl_probabilities.get_demand(index, probability, nsupply)*50
#initial calculations
sdratio = ma.filled(supply/ma.masked_where(demand==0, demand),2.0)
sdratio = _round(sdratio, 1.0, atol=CLOSE)
constrained_locations = logical_and(sdratio<1.0,demand-supply>CLOSE).astype("int8")
unconstrained_locations = 1-constrained_locations
excess_demand = (demand-supply)*constrained_locations
global_excess_demand = excess_demand.sum()
sdratio_matrix = sdratio[index]
constrained_locations_matrix = constrained_locations[index]
constrained_ex_ante = constrained_locations_matrix
# Would like to include following print statements in debug printing
# logger.log_status('Total demand:',demand.sum())
# logger.log_status('Total supply:',supply.sum())
logger.log_status('Global excess demand:',global_excess_demand)
# logger.log_status('Constrained locations:',constrained_locations.sum())
unconstrained_locations_matrix = unconstrained_locations[index]
# omega = ones(nobs,type=float32)
# pi = self.constrain_probabilities.get_pi(sdratio_matrix, omega, constrained_locations_matrix, unconstrained_locations_matrix, nobs)
omega = self.mnl_probabilities.get_omega(probability, constrained_locations_matrix, unconstrained_locations_matrix, sdratio_matrix)
omega = _round(omega, 1.0, CLOSE)
print 'Num of constrainted locations: ', constrained_locations.sum()
print 'Num of unconstrainted locations: ', unconstrained_locations.sum()
print 'Min Ex Ante Constraints:',min(constrained_ex_ante.sum(axis=1))
print 'Max Ex Ante Constraints:',max(constrained_ex_ante.sum(axis=1))
#print 'Omega shape',omega.shape
#print 'Omega histogram',histogram(omega,0,4,40)
print 'Minimum Omega',min(omega)
print 'Maximum Omega',max(omega)
print 'Mean Omega:',mean(omega)
print 'Median Omega:',median(omega)
print 'Sum Omega:',omega.sum()
print 'Standard Deviation Omega:',standard_deviation(omega)
print 'Count of Negative Omega',(where(omega<0,1,0).sum())
print 'Count of Omega < 1',(where(omega<1,1,0).sum())
print 'Count of Omega > 2',(where(omega>2,1,0).sum())
print 'Count of Omega > 4',(where(omega>4,1,0).sum())
average_omega = self.mnl_probabilities.get_average_omega(omega, probability, index, nsupply, nobs, demand)
average_omega=_round(average_omega, 1.0, CLOSE)
coef_names = resources.get("coefficient_names", None)
if coef_names is not None:
coef_names = array(coef_names.tolist()+["ln_pi"])
resources.merge({"coefficient_names":coef_names})
data=concatenate((data,ones((nobs,neqs,1),dtype=float32)), axis=2)
prev_omega = omega
prev_constrained_locations_matrix = constrained_locations_matrix
for i in range(max_iter):
print
print 'Iteration',i
pi = self.mnl_probabilities.get_pi(sdratio_matrix, omega, constrained_locations_matrix, unconstrained_locations_matrix, nobs)
#print 'pi shape',pi.shape
#print 'data shape', data.shape
#print 'min_pi',min(pi,axis=1)
#print 'max_pi',max(pi,axis=1)
#print 'min_data',min(data,axis=1)
#print 'max_data',max(data,axis=1)
data[:,:,-1] = ln(pi)
#data = concatenate((data,(pi[:,:,newaxis])),axis=-1)
#print 'data shape after contatenating pi', data.shape
result = self.bhhh_estimation.run(data, modified_upc_sequence, resources)
#print
#print 'result',result
probability = modified_upc_sequence.get_probabilities()
prob_hat = ma.filled(probability / pi, 0.0)
# HARD CODED
# scale new_demand from 0.2% to 100%
demand_new = self.mnl_probabilities.get_demand(index, prob_hat, nsupply)*50
##update supply-demand ratio
sdratio = ma.filled(supply/ma.masked_where(demand_new==0, demand_new),2.0)
sdratio = _round(sdratio, 1.0, CLOSE)
sdratio_matrix = sdratio[index]
constrained_locations = where(((average_omega*demand_new - supply) > CLOSE),1,0)
unconstrained_locations = 1-constrained_locations
constrained_locations_matrix = constrained_locations[index]
unconstrained_locations_matrix = unconstrained_locations[index]
constrained_ex_post = constrained_locations_matrix
constrained_ex_post_not_ex_ante = where((constrained_ex_post - constrained_ex_ante)==1,1,0)
constrained_ex_ante_not_ex_post = where((constrained_ex_post - constrained_ex_ante)==-1,1,0)
#Assumption 5: if j belongs to constrained ex post and unconstrained ex ante, then p^i_j <= D_j / S_j
print 'Number of individual violating Assumption 5: ', where((probability > 1 / sdratio_matrix)*constrained_ex_post_not_ex_ante)[0].size
#Assumption 6: pi of constrained locations should be less than 1
print 'Number of individual violating Assumption 6: ', where((probability * constrained_ex_post).sum(axis=1) >
(prob_hat * constrained_ex_post).sum(axis=1))[0].size
##OR ?
#print 'Assumption 6: ', where(pi[where(constrained_locations_matrix)] > 1)[0].size
print 'number of constrainted locations: ', constrained_locations.sum()
print 'number of unconstrainted locations: ', unconstrained_locations.sum()
print 'Min Ex Post Constraints:',min(constrained_ex_post.sum(axis=1))
print 'Max Ex Post Constraints:',max(constrained_ex_post.sum(axis=1))
print 'At Least 1 Constrained Ex Ante Not Ex Post*:',where(constrained_ex_ante_not_ex_post.sum(axis=1))[0].size
print 'At Least 1 Constrained Ex Post Not Ex Ante:',where(constrained_ex_post_not_ex_ante.sum(axis=1))[0].size
omega = self.mnl_probabilities.get_omega(prob_hat, constrained_locations_matrix, unconstrained_locations_matrix, sdratio_matrix)
omega = _round(omega, 1.0, CLOSE)
#print 'Omega histogram',histogram(omega,0,4,40)
print 'Minimum Omega',min(omega)
print 'Maximum Omega',max(omega)
print 'Mean Omega:',mean(omega)
print 'Median Omega:',median(omega)
print 'Sum Omega:',omega.sum()
print 'Standard Deviation Omega:',standard_deviation(omega)
print 'Count of Negative Omega',(where(omega<0,1,0).sum())
print 'Count of Omega < 1: ',(where(omega<1,1,0).sum())
print 'Count of Omega > 2: ',(where(omega>2,1,0).sum())
print 'Count of Omega > 4: ',(where(omega>4,1,0).sum())
average_omega = self.mnl_probabilities.get_average_omega(omega, prob_hat, index, nsupply, nobs, demand_new)
average_omega = _round(average_omega, 1.0, CLOSE)
excess_demand = (demand_new-supply)*constrained_locations
global_excess_demand = excess_demand.sum()
#print 'Omega [i], [i-1]',prev_omega, omega,
#print 'Constrained locations [i], [i-1]',constrained_locations_matrix, prev_constrained_locations_matrix
print 'Global Excess Demand',global_excess_demand
if ma.allclose(omega, prev_omega, atol=1e-3) or not any(constrained_locations_matrix - prev_constrained_ex_ante):
print 'omega or constrained ex post unchanged: Convergence criterion achieved'
break
#if global_excess_demand < 1:
#print 'Global excess demand < 1: Convergence criterion achieved'
#break
return result
def inner_loop(self, supply, demand, probability, index, sdratio_matrix, J,
max_iteration=100):
#vacancy_rate = SessionConfiguration().get_dataset_from_pool("vacancy_rate")
CLOSE = SessionConfiguration()["CLOSE"]
average_omega = ones(J, dtype=float32)
inner_iterations=None; constrained_locations_history = None
swing_index=array([]); average_omega_history=average_omega[:, newaxis];
N = probability.shape[0]
constrained_ex_ante = zeros(probability.shape) - 1
logger.start_block('Inner Loop')
try:
for i in range(1, max_iteration+1):
logger.log_status('Inner Loop Iteration %s' % i)
#initial calculations
constrained_locations = where(((average_omega * demand - supply) > CLOSE),1,0)
constrained_locations_matrix = constrained_locations[index]
omega = self.mnl_probabilities.get_omega(probability, constrained_locations_matrix, sdratio_matrix)
#omega = _round(omega, 1.0, CLOSE)
logger.log_status('Num of constrainted locations: %s' % constrained_locations.sum())
logger.log_status('Num of unconstrainted locations: %s' % (constrained_locations.size - constrained_locations.sum()))
logger.log_status('Min Ex Ante Constraints: %s' % min(constrained_locations_matrix.sum(axis=1)))
logger.log_status('Max Ex Ante Constraints: %s' % max(constrained_locations_matrix.sum(axis=1)))
logger.log_status('Minimum Omega: %s' % min(omega))
logger.log_status('Maximum Omega: %s' % max(omega))
logger.log_status('Mean Omega: %s' % mean(omega))
logger.log_status('Median Omega: %s' % median(omega))
logger.log_status('Sum Omega: %s' % omega.sum())
logger.log_status('Standard Deviation Omega: %s' % standard_deviation(omega))
logger.log_status('Count of Negative Omega: %s' % (where(omega<0,1,0)).sum())
logger.log_status('Count of Omega < 1: %s' % (where(omega<1,1,0)).sum())
logger.log_status('Count of Omega > 2: %s' % (where(omega>2,1,0)).sum())
logger.log_status('Count of Omega > 4: %s' % (where(omega>4,1,0)).sum())
average_omega = self.mnl_probabilities.get_average_omega(omega, probability, index, J, demand)
#average_omega=_round(average_omega, 1.0, CLOSE)
logger.log_status('Minimum average_omega: %s' % min(average_omega))
logger.log_status('Maximum average_omega: %s' % max(average_omega))
logger.log_status('Mean average_omega: %s' % mean(average_omega))
logger.log_status('Median average_omega: %s' % median(average_omega))
logger.log_status('Sum average_omega: %s' % average_omega.sum())
logger.log_status('Standard Deviation average_omega: %s' % standard_deviation(average_omega))
logger.log_status(' ')
if not any(constrained_ex_ante - constrained_locations_matrix):
return constrained_locations_matrix, omega, (inner_iterations, constrained_locations_history,swing_index, average_omega_history)
else:
constrained_ex_ante = constrained_locations_matrix
if constrained_locations_history is None:
inner_iterations = [str(i)]
constrained_locations_history = constrained_locations[:,newaxis]
else:
inner_iterations += [str(i)]
constrained_locations_history = concatenate((constrained_locations_history,
constrained_locations[:, newaxis]),
axis=1)
average_omega_history = concatenate((average_omega_history, average_omega[:, newaxis]), axis=1)
if i > 2 and ma.allclose(constrained_locations_history[:,i-1], constrained_locations_history[:,i-3]):
swing_index = where((constrained_locations_history[:,i-1] - constrained_locations_history[:,i-2]) <> 0)[0]
logger.log_warning("swing of constraints found in %s alternatives" % swing_index.size)
return constrained_locations_matrix, omega, (inner_iterations, constrained_locations_history,swing_index, average_omega_history)
finally:
logger.end_block()
logger.log_error("max iteration reached without convergence.")
raise RuntimeError, "max iteration reached without convergence."
def run(self, data, upc_sequence, resources=None):
self.mnl_probabilities=upc_sequence.probability_class
self.bhhh_estimation = bhhh_mnl_estimation()
modified_upc_sequence = UPCFactory().get_model(
utilities=None, probabilities="opus_core.mnl_probabilities", choices=None)
modified_upc_sequence.utility_class = upc_sequence.utility_class
N, neqs, V = data.shape
max_iter = resources.get("max_iterations", 100) # default
sc = SessionConfiguration()
dataset_pool = sc.get_dataset_pool()
sample_rate = dataset_pool.get_dataset("sample_rate")
CLOSE = sc["CLOSE"]
info_filename = sc["info_file"]
info_filename = os.path.join('.', info_filename)
info_file = open(info_filename, "a")
constraint_dict = {1:'constrained', 0:'unconstrained'}
swing_cases_fix = 0 #set swing alternatives to constrained (1) or unconstrained (0)
prob_correlation = None
choice_set = resources['_model_'].choice_set
J = choice_set.size()
alt_id = choice_set.get_id_attribute()
movers = choice_set.get_attribute('movers')
resources.check_obligatory_keys(["capacity_string"])
supply = choice_set.get_attribute(resources["capacity_string"])
index = resources.get("index", None)
if index is None: # no sampling case, alternative set is the full choice_set
index = arange(J)
if index.ndim <= 1:
index = repeat(index[newaxis,:], N, axis=0)
if resources.get('aggregate_to_dataset', None):
aggregate_dataset = dataset_pool.get_dataset(resources.get('aggregate_to_dataset'))
choice_set_aggregate_id = choice_set.get_attribute(aggregate_dataset.get_id_name()[0])
index = aggregate_dataset.get_id_index(choice_set_aggregate_id[index].ravel()).reshape(index.shape)
supply = aggregate_dataset.get_attribute(resources["capacity_string"])
J = aggregate_dataset.size()
movers = aggregate_dataset.get_attribute("movers")
demand_history = movers[:, newaxis]
resources.merge({"index":index})
pi = ones(index.shape, dtype=float32) #initialize pi
#average_omega = ones(J,dtype=float32) #initialize average_omega
logger.start_block('Outer Loop')
for i in range(max_iter):
logger.log_status('Outer Loop Iteration %s' % i)
result = self.bhhh_estimation.run(data, modified_upc_sequence, resources)
del self.bhhh_estimation; collect()
self.bhhh_estimation = bhhh_mnl_estimation()
probability = modified_upc_sequence.get_probabilities()
if data.shape[2] == V: #insert a placeholder for ln(pi) in data
data = concatenate((data,ones((N,neqs,1),dtype=float32)), axis=2)
coef_names = resources.get("coefficient_names")
coef_names = concatenate( (coef_names, array(["ln_pi"])) )
resources.merge({"coefficient_names":coef_names})
else:
beta_ln_pi = result['estimators'][where(coef_names == 'ln_pi')][0]
logger.log_status("mu = 1/%s = %s" % (beta_ln_pi, 1/beta_ln_pi))
prob_hat = safe_array_divide(probability, pi ** beta_ln_pi)
#prob_hat = safe_array_divide(probability, pi)
prob_hat_sum = prob_hat.sum(axis=1, dtype=float32)
if not ma.allclose(prob_hat_sum, 1.0):
logger.log_status("probability doesn't sum up to 1, with minimum %s, and maximum %s" %
(prob_hat_sum.min(), prob_hat_sum.max()))
probability = normalize(prob_hat)
demand = self.mnl_probabilities.get_demand(index, probability, J) * 1 / sample_rate
demand_history = concatenate((demand_history,
demand[:, newaxis]),
axis=1)
sdratio = safe_array_divide(supply, demand, return_value_if_denominator_is_zero=2.0)
sdratio_matrix = sdratio[index]
## debug info
from numpy import histogram
from opus_core.misc import unique
cc = histogram(index.ravel(), unique(index.ravel()))[0]
logger.log_status( "=================================================================")
logger.log_status( "Probability min: %s, max: %s" % (probability.min(), probability.max()) )
logger.log_status( "Demand min: %s, max: %s" % (demand.min(), demand.max()) )
logger.log_status( "sdratio min: %s, max: %s" % (sdratio.min(), sdratio.max()) )
logger.log_status( "demand[sdratio==sdratio.min()]=%s" % demand[sdratio==sdratio.min()] )
logger.log_status( "demand[sdratio==sdratio.max()]=%s" % demand[sdratio==sdratio.max()] )
logger.log_status( "Counts of unique submarkets in alternatives min: %s, max: %s" % (cc.min(), cc.max()) )
logger.log_status( "=================================================================")
constrained_locations_matrix, omega, info = self.inner_loop(supply, demand, probability,
index, sdratio_matrix,
J, max_iteration=max_iter)
inner_iterations, constrained_locations_history, swing_index, average_omega_history = info
for idx in swing_index:
logger.log_status("swinging alt with id %s set to %s" % (alt_id[idx], constraint_dict[swing_cases_fix]))
constrained_locations_matrix[index==idx] = swing_cases_fix
if swing_index.size > 0:
info_file.write("swing of constraints found with id %s \n" % alt_id[swing_index])
info_file.write("outer_iteration, %i, " % i + ", ".join([str(i)]*(len(inner_iterations))) + "\n")
info_file.write("inner_iteration, , " + ", ".join(inner_iterations) + "\n")
info_file.write("id, sdratio, " + ", ".join(["avg_omega"]*len(inner_iterations)) + "\n")
for idx in swing_index:
line = str(alt_id[idx]) + ','
line += str(sdratio[idx]) + ','
line += ",".join([str(x) for x in average_omega_history[idx,]])
line += "\n"
info_file.write(line)
info_file.write("\n")
info_file.flush()
outer_iterations = [str(i)] * len(inner_iterations)
prob_min = [str(probability.min())] * len(inner_iterations)
prob_max = [str(probability.max())] * len(inner_iterations)
pi_new = self.mnl_probabilities.get_pi(sdratio_matrix, omega, constrained_locations_matrix)
data[:,:,-1] = ln(pi_new)
#diagnostic output
if not ma.allclose(pi, pi_new, atol=CLOSE):
if i > 0: #don't print this for the first iteration
logger.log_status("min of abs(pi(l+1) - pi(l)): %s" % absolute(pi_new - pi).min())
logger.log_status("max of abs(pi(l+1) - pi(l)): %s" % absolute(pi_new - pi).max())
logger.log_status("mean of pi(l+1) - pi(l): %s" % (pi_new - pi).mean())
logger.log_status('Standard Deviation pi(l+1) - pi(l): %s' % standard_deviation(pi_new - pi))
logger.log_status('correlation of pi(l+1) and pi(l): %s' % corr(pi_new.ravel(), pi.ravel())[0,1])
pi = pi_new
probability_old = probability # keep probability of the previous loop, for statistics computation only
else: #convergence criterion achieved, quiting outer loop
logger.log_status("pi(l) == pi(l+1): Convergence criterion achieved")
info_file.write("\nConstrained Locations History:\n")
info_file.write("outer_iteration," + ",".join(outer_iterations) + "\n")
info_file.write("inner_iteration," + ",".join(inner_iterations) + "\n")
info_file.write("minimum_probability," + ",".join(prob_min) + "\n")
info_file.write("maximum_probability," + ",".join(prob_max) + "\n")
for row in range(J):
line = [str(x) for x in constrained_locations_history[row,]]
info_file.write(str(alt_id[row]) + "," + ",".join(line) + "\n")
info_file.flush()
info_file.write("\nDemand History:\n")
i_str = [str(x) for x in range(i)]
info_file.write("outer_iteration, (movers)," + ",".join(i_str) + "\n")
#info_file.write(", ,\n")
for row in range(J):
line = [str(x) for x in demand_history[row,]]
info_file.write(str(alt_id[row]) + "," + ",".join(line) + "\n")
demand_history_info_criteria = [500, 100, 50, 20]
for criterion in demand_history_info_criteria:
com_rows_index = where(movers <= criterion)[0]
info_file.write("\nDemand History for alternatives with less than or equal to %s movers in 1998:\n" % criterion)
i_str = [str(x) for x in range(i)]
info_file.write("outer_iteration, (movers)," + ",".join(i_str) + "\n")
#info_file.write(", movers,\n")
for row in com_rows_index:
line = [str(x) for x in demand_history[row,]]
info_file.write(str(alt_id[row]) + "," + ",".join(line) + "\n")
#import pdb; pdb.set_trace()
#export prob correlation history
correlation_indices, prob_correlation = self.compute_prob_correlation(probability_old, probability, prob_hat, index, resources)
info_file.write("\nCorrelation of Probabilities:\n")
c_name = ['corr(p_ij p~_ij)', 'corr(p_ij p^_ij)', 'corr(p_ij dummy)', 'corr(p~_ij p^_ij)', 'corr(p~_ij dummy)', 'corr(p^_ij dummy)']
info_file.write("com_id, " + ",".join(c_name) + "\n")
#info_file.write(", ,\n")
for row in range(correlation_indices.size):
line = [str(x) for x in prob_correlation[row,]]
info_file.write(str(alt_id[correlation_indices[row]]) + "," + ",".join(line) + "\n")
info_file.close()
result['pi'] = pi
return result
logger.end_block()
try:info_file.close()
except:pass
raise RuntimeError, "max iteration reached without convergence."
from urbansim.datasets.household_dataset import HouseholdDataset
from urbansim.datasets.job_dataset import JobDataset
from urbansim.models.household_location_choice_model_creator import HouseholdLocationChoiceModelCreator
from opus_core.coefficients import Coefficients
from opus_core.equation_specification import EquationSpecification
from opus_core.tests.stochastic_test_case import StochasticTestCase
from opus_core.logger import logger
niter=100
#result = run_ALCM(niter)
result = run_HTM(niter)
ngcs = result.shape[1]
var = zeros(ngcs, dtype=float32)
vart = zeros(ngcs, dtype=float32)
means = zeros(ngcs, dtype=float32)
meanst = zeros(ngcs, dtype=float32)
for ig in range(ngcs):
var[ig] = standard_deviation(result[:,ig])**2.0
means[ig] = result[:,ig].mean()
vart[ig] = standard_deviation(sqrt(result[:,ig]))**2.0
meanst[ig] = sqrt(result[:,ig]).mean()
print means
print var
print sqrt(var)
plot(means,var, 'ro')
show()
print meanst
print vart
plot(meanst,vart, 'ro')
show()