def probas(self, betas=None):
# The choice model is a logit, with availability conditions
prob1 = models.logit(self.V, self.av, 1)
prob2 = models.logit(self.V, self.av, 2)
prob3 = models.logit(self.V, self.av, 3)
simulate = {'Train': prob1, 'SM': prob2, 'Car': prob3}
biogeme = bio.BIOGEME(self.database, simulate)
biogeme.modelName = "01logit_simul"
self.biogeme.generateHtml = False
return biogeme.simulate(betas)
def logit(THE_B_TIME_RND):
"""
Calculate the conditional logit model for a given random parameter.
"""
V1 = ASC_TRAIN + \
THE_B_TIME_RND * TRAIN_TT_SCALED + \
B_COST * TRAIN_COST_SCALED
V2 = ASC_SM + \
THE_B_TIME_RND * SM_TT_SCALED + \
B_COST * SM_COST_SCALED
V3 = ASC_CAR + \
THE_B_TIME_RND * CAR_TT_SCALED + \
B_COST * CAR_CO_SCALED
# Associate utility functions with the numbering of alternatives
V = {1: V1, 2: V2, 3: V3}
# Associate the availability conditions with the alternatives
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
# The choice model is a logit, with availability conditions
integrand = models.logit(V, av, CHOICE)
return integrand
def run_simulation(data_file_directory_for_simulation, data_file_name_for_simulation, output_directory_for_simulation,
betas, household_income_limit):
"""
:author: Antonin Danalet, based on the example '01logit_simul.py' by Michel Bierlaire, EPFL, on biogeme.epfl.ch
Simulation with a binary logit model. Two alternatives: work from home at least some times, or not."""
# Read the data
df_persons = pd.read_csv(data_file_directory_for_simulation / data_file_name_for_simulation, ';')
database = db.Database('persons', df_persons)
# The following statement allows you to use the names of the variable as Python variable.
globals().update(database.variables)
# Parameters to be estimated
alternative_specific_constant = Beta('alternative_specific_constant', 0, None, None, 0)
b_no_post_school_education = Beta('b_no_post_school_education', 0, None, None, 0)
b_secondary_education = Beta('b_secondary_education', 0, None, None, 0)
b_tertiary_education = Beta('b_tertiary_education', 0, None, None, 0)
b_university = Beta('b_university', 0, None, None, 1)
b_male = Beta('b_male', 0, None, None, 0)
b_public_transport_connection_quality_na_home = Beta('b_public_transport_connection_quality_na_home',
0, None, None, 0)
b_public_transport_connection_quality_a_work = Beta('b_public_transport_connection_quality_are_a_work',
0, None, None, 1)
b_rural_work = Beta('b_rural_work', 0, None, None, 0)
b_home_work_distance = Beta('b_home_work_distance', 0, None, None, 0)
b_business_sector_agriculture = Beta('b_business_sector_agriculture', 0, None, None, 0)
b_business_sector_production = Beta('b_business_sector_production', 0, None, None, 0)
b_business_sector_wholesale = Beta('b_business_sector_wholesale', 0, None, None, 1)
b_business_sector_retail = Beta('b_business_sector_retail', 0, None, None, 0)
b_business_sector_gastronomy = Beta('b_business_sector_gastronomy', 0, None, None, 0)
b_business_sector_finance = Beta('b_business_sector_finance', 0, None, None, 1)
b_business_sector_services_fc = Beta('b_business_sector_services_fc', 0, None, None, 0)
b_business_sector_other_services = Beta('b_business_sector_other_services', 0, None, None, 1)
b_business_sector_others = Beta('b_business_sector_others', 0, None, None, 1)
b_business_sector_non_movers = Beta('b_business_sector_non_movers', 0, None, None, 0)
b_executives = Beta('b_executives', 0, None, None, 0)
b_german = Beta('b_german', 0, None, None, 0)
b_hh_income_8000_or_less = Beta('b_hh_income_8000_or_less', 0, None, None, 0)
# Definition of new variables
no_post_school_educ = education == 1
secondary_education = education == 2
tertiary_education = education == 3
university = education == 4
male = (sex == 1)
public_transport_quality_NA_home = (public_transport_connection_quality_ARE_home == 5)
public_transport_quality_A_work = (public_transport_connection_quality_ARE_work == 1)
home_work_distance = (home_work_crow_fly_distance * (home_work_crow_fly_distance >= 0.0) / 100000.0)
business_sector_agriculture = type_1 == 1
business_sector_retail = type_1 == 4
business_sector_gastronomy = type_1 == 5
business_sector_finance = type_1 == 6
business_sector_production = type_1 == 2
business_sector_wholesale = type_1 == 3
business_sector_services_fC = type_1 == 7
business_sector_other_services = type_1 == 8
business_sector_others = type_1 == 9
business_sector_non_movers = type_1 == 10
german = language == 1
nationality_switzerland = nation == 0
nationality_germany_austria = nation == 1
nationality_italy_vatican = nation == 2
nationality_france_monaco_s_marino = nation == 3
nationality_northwestern_europe = nation == 4
nationality_eastern_europe = nation == 7
hh_income_8000_or_less = hh_income < household_income_limit
executives = (0 < position_in_bus) * (position_in_bus < 19)
rural_work = urban_rural_typology_work == 3
# Utility
utility_function_telecommuting = alternative_specific_constant + \
b_executives * executives + \
b_no_post_school_education * no_post_school_educ + \
b_secondary_education * secondary_education + \
b_tertiary_education * tertiary_education + \
b_university * university + \
b_male * male + \
b_public_transport_connection_quality_na_home * public_transport_quality_NA_home + \
b_public_transport_connection_quality_a_work * public_transport_quality_A_work + \
b_rural_work * rural_work + \
b_home_work_distance * home_work_distance + \
models.piecewiseFormula(age, [0, 20, 35, 75, 200]) + \
b_business_sector_agriculture * business_sector_agriculture + \
b_business_sector_retail * business_sector_retail + \
b_business_sector_gastronomy * business_sector_gastronomy + \
b_business_sector_finance * business_sector_finance + \
b_business_sector_production * business_sector_production + \
b_business_sector_wholesale * business_sector_wholesale + \
b_business_sector_services_fc * business_sector_services_fC + \
b_business_sector_other_services * business_sector_other_services + \
b_business_sector_others * business_sector_others + \
b_business_sector_non_movers * business_sector_non_movers + \
b_german * german + \
b_nationality_ch_germany_france_italy_nw_e * nationality_switzerland + \
b_nationality_ch_germany_france_italy_nw_e * nationality_germany_austria + \
b_nationality_ch_germany_france_italy_nw_e * nationality_italy_vatican + \
b_nationality_ch_germany_france_italy_nw_e * nationality_france_monaco_s_marino + \
b_nationality_ch_germany_france_italy_nw_e * nationality_northwestern_europe + \
b_nationality_ch_germany_france_italy_nw_e * nationality_eastern_europe + \
models.piecewiseFormula(work_percentage, [0, 90, 101]) + \
b_hh_income_8000_or_less * hh_income_8000_or_less
utility_function_no_telecommuting = 0
# Associate utility functions with the numbering of alternatives
utility_functions_with_numbering_of_alternatives = {1: utility_function_telecommuting, # Yes or sometimes
3: utility_function_no_telecommuting} # No
availability_conditions = {1: 1, # Always available
3: 1} # Always available
# The choice model is a logit, with availability conditions
prob_telecommuting = models.logit(utility_functions_with_numbering_of_alternatives, availability_conditions, 1)
prob_no_telecommuting = models.logit(utility_functions_with_numbering_of_alternatives, availability_conditions, 3)
simulate = {'Prob. telecommuting': prob_telecommuting,
'Prob. no telecommuting': prob_no_telecommuting}
# Create the Biogeme object
biogeme = bio.BIOGEME(database, simulate)
biogeme.modelName = 'logit_telecommuting_simul'
# Define level of verbosity
logger = msg.bioMessage()
# logger.setSilent()
logger.setWarning()
# logger.setGeneral()
# logger.setDetailed()
# Get the betas from the estimation (without corrections)
# path_to_estimation_folder = Path('../data/output/models/estimation/')
# if os.path.isfile(path_to_estimation_folder / 'logit_telecommuting~00.pickle'):
# raise Exception('There are several model outputs! Careful.')
# results = res.bioResults(pickleFile=path_to_estimation_folder / 'logit_telecommuting.pickle')
# betas_without_correction = results.getBetaValues()
# Change the working directory, so that biogeme writes in the correct folder, i.e., where this file is
standard_directory = os.getcwd()
os.chdir(output_directory_for_simulation)
results = biogeme.simulate(theBetaValues=betas)
# print(results.describe())
df_persons = pd.concat([df_persons, results], axis=1)
# Go back to the normal working directory
os.chdir(standard_directory)
# For unemployed people, fix probability of doing some home office to 0 (and probability of not doing to 1).
df_persons.loc[df_persons.employed == 0, 'Prob. telecommuting'] = 0.0 # Unemployed people
df_persons.loc[df_persons.employed == 0, 'Prob. no telecommuting'] = 1.0 # Unemployed people
df_persons.loc[df_persons.employed == -99, 'Prob. telecommuting'] = 0.0 # Other people
df_persons.loc[df_persons.employed == -99, 'Prob. no telecommuting'] = 1.0 # Other people
# By definition, apprentices don't work from home (because they were not asked in the MTMC)
df_persons.loc[df_persons.position_in_bus == 3, 'Prob. telecommuting'] = 0.0
df_persons.loc[df_persons.position_in_bus == 3, 'Prob. no telecommuting'] = 1.0
# Add a realisation of the probability
df_persons['random 0/1'] = np.random.rand(len(df_persons))
df_persons['telecommuting_model'] = np.where(df_persons['random 0/1'] < df_persons['Prob. telecommuting'], 1, 0)
del df_persons['random 0/1']
''' Save the file '''
data_file_name = 'persons_from_SynPop_with_probability_telecommuting.csv'
df_persons.to_csv(output_directory_for_simulation / data_file_name, sep=',', index=False)
ASC_SM_RND[i] + B_TIME_RND[i] * SM_TT_SCALED + B_COST[i] * SM_COST_SCALED
for i in range(numberOfClasses)
]
V3 = [
ASC_CAR_RND[i] + B_TIME_RND[i] * CAR_TT_SCALED + B_COST[i] * CAR_CO_SCALED
for i in range(numberOfClasses)
]
V = [{1: V1[i], 2: V2[i], 3: V3[i]} for i in range(numberOfClasses)]
# Associate the availability conditions with the alternatives
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
# The choice model is a discrete mixture of logit, with availability conditions
# We calculate the conditional probability for each class
prob = [
PanelLikelihoodTrajectory(models.logit(V[i], av, CHOICE))
for i in range(numberOfClasses)
]
# Conditional to the random variables, likelihood for the individual.
probIndiv = PROB_class0 * prob[0] + PROB_class1 * prob[1]
# We integrate over the random variables using Monte-Carlo
logprob = log(MonteCarlo(probIndiv))
# Define level of verbosity
logger = msg.bioMessage()
#logger.setSilent()
#logger.setWarning()
logger.setGeneral()
#logger.setDetailed()
V1 = [ASC_1[i] + beta_shcostperdist1[i] *(sharecost/distance) + beta_shtime[i] * sharetime + SIGMA_SH_MAASRND[i] for i in range(numberOfClasses)]
V2 = [ASC_2[i] + beta_maascostperdist2[i] *(maascost/distance)+ beta_maastime1[i]*(maastime1) + beta_maastime2[i]* maastime2 +\
beta_extra[i]*(extra) *extra + SIGMA_SH_MAASRND[i] for i in range(numberOfClasses)]
V3 = [ASC_3[i] + beta_totcostperdist[i] * (currcost/distance) for i in range(numberOfClasses)]
V = [{1: V1[i],
2: V2[i],
3: V3[i]} for i in range(numberOfClasses)]
# Associate the availability conditions with the alternatives
av = {1: availability1,2: availability2,3: availability3}
# The choice model is a discrete mixture of logit, with availability conditions
# We calculate the conditional probability for each class
prob = [PanelLikelihoodTrajectory(models.logit(V[i], av, choice)) for i in range(numberOfClasses)]
# Class membership model
W = CLASS_MAAS_1 + beta_enthu_class1 * factor1 + beta_fru_class1 * factor2 + beta_edu2_1 * edu_HBO + \
beta_inc2_1*inc_mid + beta_age * age_10_60 + beta_freq * mediumfreq
PROB_class0 = models.logit({0: W, 1: 0}, None, 0)
PROB_class1 = models.logit({0: W, 1: 0}, None, 1)
# Conditional to the random variables, likelihood for the individual.
probIndiv = PROB_class0 * prob[0] + PROB_class1 * prob[1]
# We integrate over the random variables using Monte-Carlo
logprob = log(MonteCarlo(probIndiv))
# Define level of verbosity
logger = msg.bioMessage()
#logger.setSilent()
def estimate_model(self, pandas_df_for_specified_country, country):
'''
:param pandas_df_for_specified_country:
:param country:
:return: The estimated model, in a variable with 3 attributes: betas, structure, results.
'''
mypanda = pandas_df_for_specified_country
for i in range(1, 7):
mypanda['OCC_' + str(i)] = np.where(pandas_df_for_specified_country['user_occupation'] == i, 1, 0)
# create the respective database (needed for biogeme)
estimationdb = db.Database('estimationdb', mypanda)
print('Training Mode Choice model for', country)
# Alternative Specific Constants
ASC_CAR = Beta('ASC_CAR', 0, None, None, 1) # This ASC remains equal to zero
ASC_PT = Beta('ASC_PT', 0, None, None, 0)
ASC_MOT = Beta('ASC_MOT', 0, None, None, 0)
ASC_BIKE = Beta('ASC_BIKE', 0, None, None, 0)
# Beta variables (i.e. coefficients) - alternative specific
BETA_TIME = Beta('BETA_TIME', 0, None, None, 0) # Travel Time
BETA_COST = Beta('BETA_COST', 0, None, None, 0) # Travel Cost
BETA_S = Beta('BETA_S', 0, None, None, 0) # Comfort
# Beta variables (i.e. coefficients) - traveller
BETA_AGE_PT = Beta('BETA_AGE_PT', 0, None, None, 0) # Age
BETA_NCAR_PT = Beta('BETA_NCAR_PT', 0, None, None, 0) # Number of trips by car
BETA_NPT_PT = Beta('BETA_NPT_PT', 0, None, None, 0) # Number of trips by pt
BETA_GENDER_PT = Beta('BETA_GENDER_PT', 0, None, None, 0) # Gender
BETA_SCOPE_PT = Beta('BETA_SCOPE_PT', 0, None, None, 0) # Trip Purpose
BETA_OCC_1_PT = Beta('BETA_OCC_1_PT', 0, None, None, 0) # 1:Private employee
BETA_OCC_2_PT = Beta('BETA_OCC_2_PT', 0, None, None, 0) # 2:Public servant
BETA_OCC_3_PT = Beta('BETA_OCC_3_PT', 0, None, None, 0) # 3:Self-employed
BETA_OCC_5_PT = Beta('BETA_OCC_5_PT', 0, None, None, 0) # 5:Retired
BETA_OCC_6_PT = Beta('BETA_OCC_6_PT', 0, None, None, 0) # 6:Unemployed
BETA_AGE_BIKE = Beta('BETA_AGE_BIKE', 0, None, None, 0) # Age
BETA_NCAR_BIKE = Beta('BETA_NCAR_BIKE', 0, None, None, 0) # Number of trips by car
BETA_NPT_BIKE = Beta('BETA_NPT_BIKE', 0, None, None, 0) # Number of trips by pt
BETA_OCC_1_BIKE = Beta('BETA_OCC_1_BIKE', 0, None, None, 0) # 1:Private employee
BETA_OCC_3_BIKE = Beta('BETA_OCC_3_BIKE', 0, None, None, 0) # 3:Self-employed
BETA_OCC_4_BIKE = Beta('BETA_OCC_4_BIKE', 0, None, None, 0) # 4:Student
BETA_OCC_5_BIKE = Beta('BETA_OCC_5_BIKE', 0, None, None, 0) # 5:Retired
BETA_OCC_6_BIKE = Beta('BETA_OCC_6_BIKE', 0, None, None, 0) # 6:Unemployed
BETA_AGE_MOT = Beta('BETA_AGE_MOT', 0, None, None, 0) # Age
BETA_GENDER_MOT = Beta('BETA_GENDER_MOT', 0, None, None, 0) # Gender
BETA_SCOPE_MOT = Beta('BETA_SCOPE_MOT', 0, None, None, 0) # Scope
BETA_NCAR_MOT = Beta('BETA_NCAR_MOT', 0, None, None, 0) # Number of trips by car
BETA_NPT_MOT = Beta('BETA_NPT_MOT', 0, None, None, 0) # Number of trips by pt
BETA_OCC_2_MOT = Beta('BETA_OCC_2_MOT', 0, None, None, 0) # Occupation 3
BETA_OCC_3_MOT = Beta('BETA_OCC_3_MOT', 0, None, None, 0) # Occupation 3
BETA_OCC_5_MOT = Beta('BETA_OCC_5_MOT', 0, None, None, 0) # Occupation 3
BETA_OCC_6_MOT = Beta('BETA_OCC_6_MOT', 0, None, None, 0) # Occupation 6
trip_comfort_car = Variable('trip_comfort_car')
trip_comfort_moto = Variable('trip_comfort_moto')
trip_comfort_bike = Variable('trip_comfort_moto') # in the training dataset, both moto and bike are under the moto variable
trip_comfort_pt = Variable('trip_comfort_pt')
trip_cost_car = Variable('trip_cost_car')
trip_cost_moto = Variable('trip_cost_moto')
trip_cost_bike = Variable('trip_cost_moto') # in the training dataset, both moto and bike are under the moto variable
trip_cost_pt = Variable('trip_cost_pt')
trip_dur_car = Variable('trip_dur_car')
trip_dur_moto = Variable('trip_dur_moto')
trip_dur_bike = Variable('trip_dur_moto') # in the training dataset, both moto and bike are under the moto variable
trip_dur_pt = Variable('trip_dur_pt')
trip_purpose = Variable('trip_purpose')
AGE = Variable('AGE')
user_gender = Variable('user_gender')
user_trips_car = Variable('user_trips_car')
user_trips_pt = Variable('user_trips_pt')
OCC_1 = Variable('OCC_1') # 1:Private employee
OCC_2 = Variable('OCC_2') # 2:Public servant
OCC_3 = Variable('OCC_3') # 3:Self-employed
OCC_4 = Variable('OCC_4') # 4:Student
OCC_5 = Variable('OCC_5') # 5:Retired
OCC_6 = Variable('OCC_6') # 6:Unemployed
user_choice = Variable('user_choice')
user_car_avail = Variable('user_car_avail')
user_moto_avail = Variable('user_moto_avail')
user_bike_avail = Variable('user_bike_avail')
if country == 'GR' or country == 'ES': # FIXME create a separate model for ES
### Definition of utility functions - one for each alternative:
V_CAR = ASC_CAR + \
BETA_TIME * trip_dur_car + \
BETA_S * trip_comfort_car
V_PT = ASC_PT + \
BETA_TIME * trip_dur_pt+ \
BETA_S * trip_comfort_pt + \
BETA_SCOPE_PT * trip_purpose + \
BETA_AGE_PT * AGE + \
BETA_GENDER_PT * user_gender + \
BETA_NCAR_PT * user_trips_car + \
BETA_NPT_PT * user_trips_pt + \
BETA_OCC_2_PT * OCC_2 + \
BETA_OCC_5_PT * OCC_5
V_MOT = ASC_MOT + \
BETA_TIME * trip_dur_moto + \
BETA_S * trip_comfort_moto + \
BETA_SCOPE_MOT * trip_purpose + \
BETA_AGE_MOT * AGE + \
BETA_GENDER_MOT * user_gender + \
BETA_NCAR_MOT * user_trips_car + \
BETA_NPT_MOT * user_trips_pt + \
BETA_OCC_3_MOT * OCC_3 + \
BETA_OCC_6_MOT * OCC_6
# Associate the availability conditions with the alternatives. (Does not really apply on ToD modelling)
av = {1: user_car_avail,
2: 1,
3: user_moto_avail}
# Associate utility functions with the numbering of alternatives
V = {1: V_CAR,
2: V_PT,
3: V_MOT}
elif country == 'NL':
### Definition of utility functions - one for each alternative:
V_CAR = ASC_CAR + \
BETA_COST * trip_cost_car + \
BETA_TIME * trip_dur_car + \
BETA_S * trip_comfort_car
V_PT = ASC_PT + \
BETA_COST * trip_cost_pt + \
BETA_TIME * trip_dur_pt + \
BETA_S * trip_comfort_pt + \
BETA_AGE_PT * AGE + \
BETA_NCAR_PT * user_trips_car + \
BETA_NPT_PT * user_trips_pt + \
BETA_OCC_1_PT * OCC_1 + \
BETA_OCC_3_PT * OCC_3 + \
BETA_OCC_5_PT * OCC_5 + \
BETA_OCC_6_PT * OCC_6
V_BIKE = ASC_BIKE + \
BETA_COST * trip_cost_bike + \
BETA_TIME * trip_dur_bike + \
BETA_S * trip_comfort_bike + \
BETA_AGE_BIKE * AGE + \
BETA_NCAR_BIKE * user_trips_car + \
BETA_NPT_BIKE * user_trips_pt + \
BETA_OCC_1_BIKE * OCC_1 + \
BETA_OCC_3_BIKE * OCC_3 + \
BETA_OCC_4_BIKE * OCC_4 + \
BETA_OCC_5_BIKE * OCC_5 + \
BETA_OCC_6_BIKE * OCC_6
# Associate the availability conditions with the alternatives. (Does not really apply on ToD modelling)
av = {1: user_car_avail,
2: 1,
3: user_bike_avail}
# Associate utility functions with the numbering of alternatives
V = {1: V_CAR,
2: V_PT,
3: V_BIKE}
elif country == 'PT':
### Definition of utility functions - one for each alternative:
V_CAR = ASC_CAR + \
BETA_TIME * trip_dur_car + \
BETA_COST * trip_cost_car
V_PT = ASC_PT + \
BETA_TIME * trip_dur_pt + \
BETA_COST * trip_cost_pt + \
BETA_NCAR_PT * user_trips_car + \
BETA_NPT_PT * user_trips_pt + \
BETA_OCC_3_PT * OCC_3
V_MOT = ASC_MOT + \
BETA_TIME * trip_dur_moto + \
BETA_COST * trip_cost_moto + \
BETA_AGE_MOT * AGE + \
BETA_NCAR_MOT * user_trips_car + \
BETA_NPT_MOT * user_trips_pt + \
BETA_OCC_2_MOT * OCC_2 + \
BETA_OCC_3_MOT * OCC_3 + \
BETA_OCC_5_MOT * OCC_5
# Associate the availability conditions with the alternatives. (Does not really apply on ToD modelling)
av = {1: user_car_avail,
2: 1,
3: user_moto_avail}
# Associate utility functions with the numbering of alternatives
V = {1: V_CAR,
2: V_PT,
3: V_MOT}
else:
V = 1
av = 1
print('There is no model specification for ', country)
# The choice model is a log logit, with availability conditions
logprob = bioLogLogit(util=V, av=av, choice=user_choice)
biogeme = bio.BIOGEME(database=estimationdb, formulas=logprob)
biogeme.modelName = "logitEstimation"
# Create the outputs of the estimation and store in a namedtuple (= Model)
results = biogeme.estimate()
betas = results.getBetaValues() # To be used later for the simulation of the model
structure = {1: models.logit(V, av, 1),
2: models.logit(V, av, 2),
3: models.logit(V, av, 3)}
Output = collections.namedtuple('Output', ['betas', 'structure', 'results'])
Model = Output(betas, structure, results)
self.__cleanup_after_model_training()
# print(self.evaluate_model(pandas_df_for_specified_country, Model))
return Model
# Associate utility functions with the numbering of alternatives
V = {1: V1,
2: V2,
3: V3}
# Associate the availability conditions with the alternatives
CAR_AV_SP = DefineVariable('CAR_AV_SP',CAR_AV * ( SP != 0 ),database)
TRAIN_AV_SP = DefineVariable('TRAIN_AV_SP',TRAIN_AV * ( SP != 0 ),database)
av = {1: TRAIN_AV_SP,
2: SM_AV,
3: CAR_AV_SP}
# The choice model is a logit, with availability conditions
prob1 = Elem({0:0,1:models.logit(V,av,1)},av[1])
# Elasticities can be computed. We illustrate below two
# formulas. Check in the output file that they produce the same
# result.
# First, the general definition of elasticities. This illustrates the
# use of the Derive expression, and can be used with any model,
# however complicated it is. Note the quotes in the Derive opertor.
genelas1 = Derive(prob1,'TRAIN_TT') * TRAIN_TT / prob1
# Second, the elasticity of logit models. See Ben-Akiva and Lerman for
# the formula
logitelas1 = TRAIN_AV_SP * (1.0 - prob1) * TRAIN_TT_SCALED * B_TIME
BETA_TIME_CAR = BETA_TIME_CAR_REF * exp(BETA_TIME_CAR_CL * CARLOVERS)
V1 = ASC_CAR + \
BETA_TIME_CAR * TimeCar_scaled + \
BETA_COST_HWH * CostCarCHF_scaled * PurpHWH + \
BETA_COST_OTHER * CostCarCHF_scaled * PurpOther + \
ec_sigma * errorComponent
V2 = ASC_SM + BETA_DIST * distance_km_scaled
# Associate utility functions with the numbering of alternatives
V = {0: V0, 1: V1, 2: V2}
# Conditional to the random parameters, we have a logit model (called
# the kernel) for the choice
condprob = models.logit(V, None, Choice)
# Conditional to the random parameters, we have the product of ordered
# probit for the indicators.
condlike = P_Envir01 * \
P_Envir02 * \
P_Envir03 * \
P_Mobil11 * \
P_Mobil14 * \
P_Mobil16 * \
P_Mobil17 * \
condprob
# We integrate over omega using Monte-Carlo integration
loglike = log(MonteCarlo(condlike))
CAR_TT_SCALED = CAR_TT / 100
CAR_CO_SCALED = CAR_CO / 100
# Definition of the utility functions
V1 = ASC_TRAIN + \
B_TIME_RND * TRAIN_TT_SCALED + \
B_COST * TRAIN_COST_SCALED
V2 = ASC_SM + \
B_TIME_RND * SM_TT_SCALED + \
B_COST * SM_COST_SCALED
V3 = ASC_CAR + \
B_TIME_RND * CAR_TT_SCALED + \
B_COST * CAR_CO_SCALED
# Associate utility functions with the numbering of alternatives
V = {1: V1, 2: V2, 3: V3}
# Associate the availability conditions with the alternatives
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
# The choice model is a logit, with availability conditions
integrand = models.logit(V, av, CHOICE)
numericalI = Integrate(integrand * density, 'omega')
simulate = {'Numerical': numericalI}
biogeme = bio.BIOGEME(database, simulate)
results = biogeme.simulate()
print('Mixture of logit - numerical integration: ',
results.iloc[0]['Numerical'])
ASC_SM_RND[i] + B_TIME_RND[i] * SM_TT_SCALED + B_COST[i] * SM_COST_SCALED
for i in range(numberOfClasses)
]
V3 = [
ASC_CAR_RND[i] + B_TIME_RND[i] * CAR_TT_SCALED + B_COST[i] * CAR_CO_SCALED
for i in range(numberOfClasses)
]
V = [{1: V1[i], 2: V2[i], 3: V3[i]} for i in range(numberOfClasses)]
# Associate the availability conditions with the alternatives
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
# The choice model is a discrete mixture of logit, with availability conditions
# We calculate the conditional probability for each class
prob = [
PanelLikelihoodTrajectory(models.logit(V[i], av, CHOICE))
for i in range(numberOfClasses)
]
# Class membership model
W = CLASS_CTE + CLASS_INC * INCOME
PROB_class0 = models.logit({0: W, 1: 0}, None, 0)
PROB_class1 = models.logit({0: W, 1: 0}, None, 1)
# Conditional to the random variables, likelihood for the individual.
probIndiv = PROB_class0 * prob[0] + PROB_class1 * prob[1]
# We integrate over the random variables using Monte-Carlo
logprob = log(MonteCarlo(probIndiv))
# Define level of verbosity
CAR_AV_SP = DefineVariable('CAR_AV_SP',CAR_AV * ( SP != 0 ),database)
TRAIN_AV_SP = DefineVariable('TRAIN_AV_SP',TRAIN_AV * ( SP != 0 ),database)
av = {1: TRAIN_AV_SP,
2: SM_AV,
3: CAR_AV_SP}
# Class membership model
W_OTHER = Beta('W_OTHER',0.5,0,1,0)
probClass1 = 1 - W_OTHER
probClass2 = W_OTHER
# The choice model is a discrete mixture of logit, with availability conditions
prob1 = models.logit(V1,av,CHOICE)
prob2 = models.logit(V2,av,CHOICE)
prob = probClass1 * prob1 + probClass2 * prob2
logprob = log(prob)
class test_07(unittest.TestCase):
def testEstimation(self):
biogeme = bio.BIOGEME(database,logprob)
results = biogeme.estimate()
self.assertAlmostEqual(results.data.logLike,-5208.4980304812725,2)
if __name__ == '__main__':
unittest.main()
def apply_model_to_example(df_persons, betas, output_directory_for_simulation,
output_file_name):
"""
:author: Antonin Danalet, based on the example '01logit_simul.py' by Michel Bierlaire, EPFL, on biogeme.epfl.ch
Simulation with a binary logit model. Two alternatives: work from home at least some times, or not."""
# Read the data
database = db.Database('persons', df_persons)
# The following statement allows you to use the names of the variable as Python variable.
globals().update(database.variables)
# Parameters to be estimated
alternative_specific_constant = Beta('alternative_specific_constant', 0,
None, None, 0)
b_no_post_school_education = Beta('b_no_post_school_education', 0, None,
None, 0)
b_secondary_education = Beta('b_secondary_education', 0, None, None, 0)
b_tertiary_education = Beta('b_tertiary_education', 0, None, None, 0)
b_university = Beta('b_university', 0, None, None, 1)
b_male = Beta('b_male', 0, None, None, 0)
b_public_transport_connection_quality_are_a_home = Beta(
'b_public_transport_connection_quality_are_a_home', 0, None, None, 1)
b_public_transport_connection_quality_are_b_home = Beta(
'b_public_transport_connection_quality_are_b_home', 0, None, None, 1)
b_public_transport_connection_quality_are_c_home = Beta(
'b_public_transport_connection_quality_are_c_home', 0, None, None, 1)
b_public_transport_connection_quality_are_d_home = Beta(
'b_public_transport_connection_quality_are_d_home', 0, None, None, 1)
b_public_transport_connection_quality_are_na_home = Beta(
'b_public_transport_connection_quality_are_na_home', 0, None, None, 0)
b_urban_work = Beta('b_urban_work', 0, None, None, 1)
b_rural_work = Beta('b_rural_work', 0, None, None, 0)
b_intermediate_work = Beta('b_intermediate_work', 0, None, None, 1)
b_home_work_distance = Beta('b_home_work_distance', 0, None, None, 0)
b_business_sector_agriculture = Beta('b_business_sector_agriculture', 0,
None, None, 0)
b_business_sector_production = Beta('b_business_sector_production', 0,
None, None, 0)
b_business_sector_wholesale = Beta('b_business_sector_wholesale', 0, None,
None, 1)
b_business_sector_retail = Beta('b_business_sector_retail', 0, None, None,
0)
b_business_sector_gastronomy = Beta('b_business_sector_gastronomy', 0,
None, None, 0)
b_business_sector_finance = Beta('b_business_sector_finance', 0, None,
None, 1)
b_business_sector_services_fc = Beta('b_business_sector_services_fc', 0,
None, None, 0)
b_business_sector_other_services = Beta('b_business_sector_other_services',
0, None, None, 1)
b_business_sector_others = Beta('b_business_sector_others', 0, None, None,
1)
b_business_sector_non_movers = Beta('b_business_sector_non_movers', 0,
None, None, 0)
b_employees = Beta('b_employees', 0, None, None, 1)
b_executives = Beta('b_executives', 0, None, None, 0)
b_german = Beta('b_german', 0, None, None, 0)
b_nationality_ch_germany_france_italy_nw_e = Beta(
'b_nationality_ch_germany_france_italy_nw_e', 0, None, None, 1)
b_nationality_south_west_europe = Beta('b_nationality_south_west_europe',
0, None, None, 1)
b_nationality_southeast_europe = Beta('b_nationality_southeast_europe', 0,
None, None, 1)
b_hh_income_na = Beta('B_hh_income_na', 0, None, None, 1)
b_hh_income_8000_or_less = Beta('b_hh_income_8000_or_less', 0, None, None,
0)
b_hh_income_more_than_8000 = Beta('b_hh_income_more_than_8000', 0, None,
None, 1)
# Definition of new variables
no_post_school_educ = ((highest_educ == 1) | (highest_educ == 2) |
(highest_educ == 3) | (highest_educ == 4))
secondary_education = ((highest_educ == 5) | (highest_educ == 6) |
(highest_educ == 7) | (highest_educ == 8) |
(highest_educ == 9) | (highest_educ == 10) |
(highest_educ == 11) | (highest_educ == 12))
tertiary_education = ((highest_educ == 13) | (highest_educ == 14) |
(highest_educ == 15) | (highest_educ == 16))
university = (highest_educ == 17)
male = (sex == 1)
public_transport_connection_quality_ARE_A_home = (
public_transport_connection_quality_ARE_home == 1)
public_transport_connection_quality_ARE_B_home = (
public_transport_connection_quality_ARE_home == 2)
public_transport_connection_quality_ARE_C_home = (
public_transport_connection_quality_ARE_home == 3)
public_transport_connection_quality_ARE_D_home = (
public_transport_connection_quality_ARE_home == 4)
public_transport_connection_quality_ARE_NA_home = (
public_transport_connection_quality_ARE_home == 5)
urban_work = (urban_typology_work == 1)
rural_work = (urban_typology_work == 3)
intermediate_work = (urban_typology_work == 2)
home_work_distance = (home_work_crow_fly_distance *
(home_work_crow_fly_distance >= 0.0) / 100000.0)
business_sector_agriculture = DefineVariable('business_sector_agriculture',
1 <= noga_08 <= 7, database)
business_sector_retail = DefineVariable('business_sector_retail',
47 <= noga_08 <= 47, database)
business_sector_gastronomy = DefineVariable('business_sector_gastronomy',
55 <= noga_08 <= 57, database)
business_sector_finance = DefineVariable('business_sector_finance',
64 <= noga_08 <= 67, database)
business_sector_production = DefineVariable(
'business_sector_production',
(10 <= noga_08 <= 35) | (40 <= noga_08 <= 44), database)
business_sector_wholesale = DefineVariable('business_sector_wholesale',
(45 <= noga_08 <= 45) |
(49 <= noga_08 <= 54), database)
business_sector_services_fC = DefineVariable(
'business_sector_services_fC',
(60 <= noga_08 <= 63) | (69 <= noga_08 <= 83) | (noga_08 == 58),
database)
business_sector_other_services = DefineVariable(
'business_sector_other_services', (86 <= noga_08 <= 90) |
(92 <= noga_08 <= 96) | (noga_08 == 59) | (noga_08 == 68), database)
business_sector_others = DefineVariable('business_sector_others',
97 <= noga_08 <= 98, database)
business_sector_non_movers = DefineVariable(
'business_sector_non_movers',
(8 <= noga_08 <= 9) | (36 <= noga_08 <= 39) | (84 <= noga_08 <= 85) |
(noga_08 == 91) | (noga_08 == 99), database)
employees = work_position == 2
executives = work_position == 1
german = language == 1
nationality_switzerland = nation == 8100
nationality_germany_austria_lichtenstein = (nation == 8207) + (
nation == 8229) + (nation == 8222)
nationality_italy_vatican = (nation == 8218) + (nation == 8241)
nationality_france_monaco_san_marino = (nation == 8212) + (
nation == 8226) + (nation == 8233)
nationality_northwestern_europe = (nation == 8204) + (nation == 8223) + (nation == 8227) + (nation == 8206) + \
(nation == 8211) + (nation == 8215) + (nation == 8216) + (nation == 8217) + \
(nation == 8228) + (nation == 8234)
nationality_south_west_europe = (nation == 8231) + (nation == 8236) + (
nation == 8202)
nationality_southeast_europe = (nation == 8224) + (nation == 8201) + (nation == 8214) + (nation == 8256) + \
(nation == 8250) + (nation == 8251) + (nation == 8252) + (nation == 8255) + \
(nation == 8205) + (nation == 8239) + (nation == 8242) + (nation == 8248) + \
(nation == 8254)
nationality_eastern_europe = (nation == 8230) + (nation == 8232) + (nation == 8240) + (nation == 8243) + \
(nation == 8244) + (nation == 8263) + (nation == 8265) + (nation == 8266) + \
(nation == 8260) + (nation == 8261) + (nation == 8262)
# several_part_time_jobs = full_part_time_job == 3
work_percentage = DefineVariable(
'work_percentage',
bioMin(
(full_part_time_job == 1) * 100 + percentage_first_part_time_job *
(percentage_first_part_time_job > 0), # +
# percentage_second_part_time_job * (percentage_second_part_time_job > 0),
100),
database)
hh_income_na = hh_income == -98
hh_income_less_than_2000 = hh_income == 1
hh_income_2000_to_4000 = hh_income == 2
hh_income_4001_to_6000 = hh_income == 3
hh_income_6001_to_8000 = hh_income == 4
hh_income_8001_to_10000 = hh_income == 5
hh_income_10001_to_12000 = hh_income == 6
hh_income_12001_to_14000 = hh_income == 7
hh_income_14001_to_16000 = hh_income == 8
hh_income_more_than_16000 = hh_income == 9
# Utility
U = alternative_specific_constant + \
b_executives * executives + \
b_employees * employees + \
b_no_post_school_education * no_post_school_educ + \
b_secondary_education * secondary_education + \
b_tertiary_education * tertiary_education + \
b_university * university + \
b_male * male + \
b_public_transport_connection_quality_are_a_home * public_transport_connection_quality_ARE_A_home + \
b_public_transport_connection_quality_are_b_home * public_transport_connection_quality_ARE_B_home + \
b_public_transport_connection_quality_are_c_home * public_transport_connection_quality_ARE_C_home + \
b_public_transport_connection_quality_are_d_home * public_transport_connection_quality_ARE_D_home + \
b_public_transport_connection_quality_are_na_home * public_transport_connection_quality_ARE_NA_home + \
b_urban_work * urban_work + \
b_rural_work * rural_work + \
b_intermediate_work * intermediate_work + \
b_home_work_distance * home_work_distance + \
models.piecewiseFormula(age, [0, 20, 35, 75, 200]) + \
b_business_sector_agriculture * business_sector_agriculture + \
b_business_sector_retail * business_sector_retail + \
b_business_sector_gastronomy * business_sector_gastronomy + \
b_business_sector_finance * business_sector_finance + \
b_business_sector_production * business_sector_production + \
b_business_sector_wholesale * business_sector_wholesale + \
b_business_sector_services_fc * business_sector_services_fC + \
b_business_sector_other_services * business_sector_other_services + \
b_business_sector_others * business_sector_others + \
b_business_sector_non_movers * business_sector_non_movers + \
b_german * german + \
b_nationality_ch_germany_france_italy_nw_e * nationality_switzerland + \
b_nationality_ch_germany_france_italy_nw_e * nationality_germany_austria_lichtenstein + \
b_nationality_ch_germany_france_italy_nw_e * nationality_italy_vatican + \
b_nationality_ch_germany_france_italy_nw_e * nationality_france_monaco_san_marino + \
b_nationality_ch_germany_france_italy_nw_e * nationality_northwestern_europe + \
b_nationality_south_west_europe * nationality_south_west_europe + \
b_nationality_southeast_europe * nationality_southeast_europe + \
b_nationality_ch_germany_france_italy_nw_e * nationality_eastern_europe + \
models.piecewiseFormula(work_percentage, [0, 90, 101]) + \
b_hh_income_na * hh_income_na + \
b_hh_income_8000_or_less * hh_income_less_than_2000 + \
b_hh_income_8000_or_less * hh_income_2000_to_4000 + \
b_hh_income_8000_or_less * hh_income_4001_to_6000 + \
b_hh_income_8000_or_less * hh_income_6001_to_8000 + \
b_hh_income_more_than_8000 * hh_income_8001_to_10000 + \
b_hh_income_more_than_8000 * hh_income_10001_to_12000 + \
b_hh_income_more_than_8000 * hh_income_12001_to_14000 + \
b_hh_income_more_than_8000 * hh_income_14001_to_16000 + \
b_hh_income_more_than_8000 * hh_income_more_than_16000
U_No_telecommuting = 0
# Associate utility functions with the numbering of alternatives
V = {
1: U, # Yes or sometimes
0: U_No_telecommuting
} # No
av = {1: 1, 0: 1}
# The choice model is a logit, with availability conditions
prob_telecommuting = models.logit(V, av, 1)
prob_no_telecommuting = models.logit(V, av, 0)
simulate = {
'Prob. telecommuting': prob_telecommuting,
'Prob. no telecommuting': prob_no_telecommuting
}
# Create the Biogeme object
biogeme = bio.BIOGEME(database, simulate)
biogeme.modelName = 'logit_telecommuting_simul'
# Change the working directory, so that biogeme writes in the correct folder, i.e., where this file is
# standard_directory = os.getcwd()
# os.chdir(output_directory_for_simulation)
results = biogeme.simulate(theBetaValues=betas)
# print(results.describe())
df_persons = pd.concat([df_persons, results], axis=1)
# Go back to the normal working directory
# os.chdir(standard_directory)
''' Save the file '''
df_persons.to_csv(output_directory_for_simulation / output_file_name,
sep=',',
index=False)
b_emipp = u_emipp + sd_emipp * bioDraws('b_emipp', 'NORMAL')
V1 = price_1*b_price+time_1*b_time+conven_1*b_conven+comfort_1*b_comfort+\
meals_1*b_meals+petfr_1*b_petfr+emipp_1*b_emipp+nonsig1_1*b_nonsig1+\
nonsig2_1*b_nonsig2+nonsig3_1*b_nonsig3
V2 = price_2*b_price+time_2*b_time+conven_2*b_conven+comfort_2*b_comfort+\
meals_2*b_meals+petfr_2*b_petfr+emipp_2*b_emipp+nonsig1_2*b_nonsig1+\
nonsig2_2*b_nonsig2+nonsig3_2*b_nonsig3
V3 = price_3*b_price+time_3*b_time+conven_3*b_conven+comfort_3*b_comfort+\
meals_3*b_meals+petfr_3*b_petfr+emipp_3*b_emipp+nonsig1_3*b_nonsig1+\
nonsig2_3*b_nonsig2+nonsig3_3*b_nonsig3
V = {1: V1, 2: V2, 3: V3}
av = {1: aval_1, 2: aval_2, 3: aval_3}
prob = models.logit(V, av, choice)
logprob = log(MonteCarlo(prob))
# Define level of verbosity
logger = msg.bioMessage()
logger.setSilent()
# Create the Biogeme object
biogeme = bio.BIOGEME(database,
logprob,
numberOfDraws=n_draws,
numberOfThreads=n_cores)
biogeme.modelName = 'MixedLogitArtificial'
biogeme.generateHtml = False
biogeme.generatePickle = False
# Estimate the parameters
# For latent class 2, whete the time coefficient is estimated
V21 = ASC_TRAIN + B_TIME * TRAIN_TT_SCALED + B_COST * TRAIN_COST_SCALED + EC_TRAIN
V22 = ASC_SM + B_TIME * SM_TT_SCALED + B_COST * SM_COST_SCALED + EC_SM
V23 = ASC_CAR + B_TIME * CAR_TT_SCALED + B_COST * CAR_CO_SCALED + EC_CAR
V2 = {1: V21, 2: V22, 3: V23}
# Associate the availability conditions with the alternatives
CAR_AV_SP = DefineVariable('CAR_AV_SP', CAR_AV * (SP != 0), database)
TRAIN_AV_SP = DefineVariable('TRAIN_AV_SP', TRAIN_AV * (SP != 0), database)
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
# Class membership model
W_OTHER = Beta('W_OTHER', 0.798, 0, 1, 0)
probClass1 = 1 - W_OTHER
probClass2 = W_OTHER
# The choice model is a discrete mixture of logit, with availability conditions
prob1 = PanelLikelihoodTrajectory(models.logit(V1, av, CHOICE))
prob2 = PanelLikelihoodTrajectory(models.logit(V2, av, CHOICE))
probIndiv = probClass1 * prob1 + probClass2 * prob2
logprob = log(MonteCarlo(probIndiv))
biogeme = bio.BIOGEME(database, logprob)
biogeme.modelName = "15panelDiscrete"
results = biogeme.estimate()
print("Results=", results)
V3 = [
ASC_3[i] + beta_parktime3[i] * (parktime) + beta_parkcost3[i] *
(parkcost) / 10 + beta_dist3[i] * sharedist / 10
for i in range(numberOfClasses)
]
V4 = [ASC_4[i] for i in range(numberOfClasses)]
V = [{1: V1[i], 2: V2[i], 3: V3[i], 4: V4[i]} for i in range(numberOfClasses)]
# Associate the availability conditions with the alternatives
av = {1: availability1, 2: availability2, 3: availability3, 4: availability4}
# The choice model is a discrete mixture of logit, with availability conditions
# We calculate the conditional probability for each class
prob = [
PanelLikelihoodTrajectory(models.logit(V[i], av, choice))
for i in range(numberOfClasses)
]
# Class membership model
W = CLASS_MAAS + beta_enthu * factor1 + beta_fru * factor2 + beta_inc2*inc_mid +\
beta_edu2 * edu_HBO + beta_fam * fam_1 + beta_fam * fam_2 + beta_age * age_10_60
PROB_class0 = models.logit({1: W, 0: 0}, None, 1)
PROB_class1 = models.logit({1: W, 0: 0}, None, 0)
# Conditional to the random variables, likelihood for the individual.
probIndiv = PROB_class0 * prob[0] + PROB_class1 * prob[1]
# We integrate over the random variables using Monte-Carlo
logprob = log(MonteCarlo(probIndiv))
# Define level of verbosity
B_COST * SM_COST_SCALED
V3 = ASC_CAR + \
B_TIME * CAR_TT_SCALED + \
B_COST * CAR_CO_SCALED
# Associate utility functions with the numbering of alternatives
V = {1: V1, 2: V2, 3: V3}
# Associate the availability conditions with the alternatives
CAR_AV_SP = DefineVariable('CAR_AV_SP', CAR_AV * (SP != 0), database)
TRAIN_AV_SP = DefineVariable('TRAIN_AV_SP', TRAIN_AV * (SP != 0), database)
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
# The choice model is a logit, with availability conditions
prob1 = models.logit(V, av, 1)
prob2 = models.logit(V, av, 2)
prob3 = models.logit(V, av, 3)
# Elasticities can be computed. We illustrate below two
# formulas. Check in the output file that they produce the same
# result.
# First, the general definition of elasticities. This illustrates the
# use of the Derive expression, and can be used with any model,
# however complicated it is. Note the quotes in the Derive opertor.
genelas1 = Derive(prob1, 'TRAIN_TT') * TRAIN_TT / prob1
genelas2 = Derive(prob2, 'SM_TT') * SM_TT / prob2
genelas3 = Derive(prob3, 'CAR_TT') * CAR_TT / prob3
V2 = [ASC_SM_RND[i] + B_TIME_RND[i] * SM_TT_SCALED + B_COST[i] * SM_COST_SCALED
for i in range(numberOfClasses)]
V3 = [ASC_CAR_RND[i] + B_TIME_RND[i] * CAR_TT_SCALED + B_COST[i] * CAR_CO_SCALED
for i in range(numberOfClasses)]
V = [{1: V1[i],
2: V2[i],
3: V3[i]} for i in range(numberOfClasses)]
# Associate the availability conditions with the alternatives
av = {1: TRAIN_AV_SP,
2: SM_AV,
3: CAR_AV_SP}
# The choice model is a discrete mixture of logit, with availability conditions
# We calculate the conditional probability for each class
prob = [MonteCarlo(PanelLikelihoodTrajectory(models.logit(V[i], av, CHOICE)))
for i in range(numberOfClasses)]
# Conditional to the random variables, likelihood for the individual.
probIndiv = PROB_class0 * prob[0] + PROB_class1 * prob[1]
# We integrate over the random variables using Monte-Carlo
logprob = log(probIndiv)
# Define level of verbosity
logger = msg.bioMessage()
#logger.setSilent()
#logger.setWarning()
logger.setGeneral()
#logger.setDetailed()
MaaS = ASC_2 + beta_maascostperdist2 *(maascost*10/distance)+ beta_nonebike_accesstime*(maastime1) + beta_ebike_accesstime* maastime2 +\
beta_extra*(extra) *extra +\
beta_enthu * factor1 + beta_fru * factor2 + beta_constructive * factor3 + beta_travelzeal * factor4 + beta_age_package * age_10_60 +\
beta_edu2_package * edu_WO +beta_edu2_package * edu_HBO + beta_inc2_package*inc_mid + SIGMA_SH_MAASRND + beta_freq * highfreq + beta_freq * mediumfreq
Continue_following_existing_way = ASC_3 + beta_totcostperdist * (currcost*10/distance) + beta_age_private * age_60_abv + \
beta_edu2_private * edu_WO + beta_edu2_private * edu_HBO + beta_inc2_private*inc_mid
# Associate utility functions with the numbering of alternatives
choiceset = {1: Shared_vehicle, 2: MaaS, 3: Continue_following_existing_way}
availability = {1: availability1, 2: availability2, 3: availability3}
# Definition of the model. This is the contribution of each
# observation to the log likelihood function.
# The choice model is a nested logit, with availability conditions
obsprob = models.logit(choiceset, availability, choice)
condprobIndiv = PanelLikelihoodTrajectory(obsprob)
# We integrate over the random parameters using Monte-Carlo
logprob = log(MonteCarlo(condprobIndiv))
# Define level of verbosity
logger = msg.bioMessage()
#logger.setSilent()
#logger.setWarning()
#logger.setGeneral()
logger.setDetailed()
# Create the Biogeme object
biogeme = bio.BIOGEME(database, logprob, numberOfDraws=50)
biogeme.modelName = 'Mixed logit outside'
V1 = ASC_CAR + \
BETA_TIME_CAR * TimeCar_scaled + \
BETA_COST_HWH * CostCarCHF_scaled * PurpHWH + \
BETA_COST_OTHER * CostCarCHF_scaled * PurpOther
V2 = ASC_SM + BETA_DIST * distance_km_scaled
# Associate utility functions with the numbering of alternatives
V = {0: V0, 1: V1, 2: V2}
# Associate the availability conditions with the alternatives.
# In this example all alternatives are available for each individual.
av = {0: 1, 1: 1, 2: 1}
# Conditional to omega, we have a logit model (called the kernel) for the choice
condprob = models.logit(V, av, Choice)
# Conditional to omega, we have the product of ordered probit for the indicators.
condlike = P_Envir01 * \
P_Envir02 * \
P_Envir03 * \
P_Mobil11 * \
P_Mobil14 * \
P_Mobil16 * \
P_Mobil17 * \
condprob
# We integrate over omega using numerical integration
loglike = log(Integrate(condlike * density, 'omega'))
# Define level of verbosity
V2 = ASC_SM + \
B_TIME_RND * SM_TT_SCALED + \
B_COST * SM_COST_SCALED
V3 = ASC_CAR + \
B_TIME_RND * CAR_TT_SCALED + \
B_COST * CAR_CO_SCALED
# Associate utility functions with the numbering of alternatives
V = {1: V1, 2: V2, 3: V3}
# Associate the availability conditions with the alternatives
CAR_AV_SP = DefineVariable('CAR_AV_SP', CAR_AV * (SP != 0), database)
TRAIN_AV_SP = DefineVariable('TRAIN_AV_SP', TRAIN_AV * (SP != 0), database)
av = {1: TRAIN_AV_SP, 2: SM_AV, 3: CAR_AV_SP}
obsprob = models.logit(V, av, CHOICE)
condprobIndiv = PanelLikelihoodTrajectory(obsprob)
logprob = log(MonteCarlo(condprobIndiv))
class test_12(unittest.TestCase):
def testEstimation(self):
biogeme = bio.BIOGEME(database, logprob, numberOfDraws=5, seed=10)
results = biogeme.estimate()
self.assertAlmostEqual(results.data.logLike, -4705.638407401872, 2)
if __name__ == '__main__':
unittest.main()
def estimate_model(self, pandas_df_for_specified_country, country):
'''
:param pandas_df_for_specified_country:
:param country:
:return: The estimated model, in a variable with 3 attributes: betas, structure, results.
'''
# create the respective database (needed for biogeme)
estimationdb = db.Database('estimationdb',
pandas_df_for_specified_country)
print('Training Time of Departure model for', country)
# Alternative Specific Constants
ASC_EARLIER = Beta('ASC_EARLIER', 0, None, None,
1) # This ASC remains equal to zero
ASC_ONTIME = Beta('ASC_ONTIME', 0, None, None, 0)
ASC_LATER = Beta('ASC_LATER', 0, None, None, 0)
# Beta variables (i.e. coefficients) - alternative specific
BETA_TT = Beta(
'BETA_TT', 0, None, None,
0) # Travel Time - Beta of TT1, TT2, TT3 - same across all Alts
BETA_WT = Beta(
'BETA_WT', 0, None, None,
0) # Walking Time - Beta of WT1, WT2, WT3 - same across all Alts
BETA_FREQ = Beta(
'BETA_FREQ', 0, None, None,
0) # Frequency - Beta of F1, F2, F3 - same across all Alts
BETA_FARE_DISCOUNT = Beta(
'BETA_FARE_DISCOUNT', 0, None, None,
0) # Fare Discount - Beta of C1, C2, C3 - same across all Alts
# Beta variables (i.e. coefficients) - traveller
BETA_GENDER_ONTIME = Beta('BETA_GENDER_ONTIME', 0, None, None,
0) # Beta of GENDER, for Alt 2
BETA_IMPORTANT_ONTIME = Beta('BETA_IMPORTANT_ONTIME', 0, None, None,
0) # Beta of TWORK, for Alt 2
BETA_SCOPE_ONTIME = Beta('BETA_SCOPE_ONTIME', 0, None, None,
0) # Beta of SCOPE, for Alt 2
BETA_NPT_ONTIME = Beta('BETA_NPT_ONTIME', 0, None, None,
0) # Beta of NPT, for Alt 2
BETA_GENDER_LATER = Beta('BETA_GENDER_LATER', 0, None, None,
0) # Beta of GENDER, for Alt 3
BETA_IMPORTANT_LATER = Beta('BETA_IMPORTANT_LATER', 0, None, None,
0) # Beta of TWORK, for Alt 3
BETA_SCOPE_LATER = Beta('BETA_SCOPE_LATER', 0, None, None,
0) # Beta of SCOPE, for Alt 3
BETA_NPT_LATER = Beta('BETA_NPT_LATER', 0, None, None,
0) # Beta of NPT, for Alt 3
BETA_HOUSEHOLD_LATER = Beta('BETA_HOUSEHOLD_LATER', 0, None, None,
0) # Beta of HOUSEHOLD, for Alt 3
BETA_INCOME_LATER = Beta('BETA_INCOME_LATER', 0, None, None,
0) # Beta of INCOME, for Alt 3
BETA_AGE_LATER = Beta('BETA_AGE_LATER', 0, None, None,
0) # Beta of AGE, for Alt 3
AGE = Variable('AGE')
user_income = Variable('user_income')
user_household = Variable('user_household')
user_trips_pt = Variable('user_trips_pt')
trip_discount_earlier = Variable('trip_discount_earlier')
trip_discount_later = Variable('trip_discount_later')
trip_discount_ontime = Variable('trip_discount_ontime')
trip_dur_earlier = Variable('trip_dur_earlier')
trip_dur_later = Variable('trip_dur_later')
trip_dur_ontime = Variable('trip_dur_ontime')
trip_freq_earlier = Variable('trip_freq_earlier')
trip_freq_later = Variable('trip_freq_later')
trip_freq_ontime = Variable('trip_freq_ontime')
trip_purpose = Variable('trip_purpose')
trip_walk_earlier = Variable('trip_walk_earlier')
trip_walk_later = Variable('trip_walk_later')
trip_walk_ontime = Variable('trip_walk_ontime')
user_choice = Variable('user_choice')
user_gender = Variable('user_gender')
user_imp_arr = Variable('user_imp_arr')
if country == 'GR' or country == 'ES': # FIXME create a separate model for ES
# Definition of utility functions - one for each alternative:
V_EARLIER = ASC_EARLIER + \
BETA_TT * trip_dur_earlier + \
BETA_WT * trip_walk_earlier + \
BETA_FREQ * trip_freq_earlier
V_ONTIME = ASC_ONTIME + \
BETA_TT * trip_dur_ontime + \
BETA_WT * trip_walk_ontime + \
BETA_FREQ * trip_freq_ontime + \
BETA_IMPORTANT_ONTIME * user_imp_arr + \
BETA_NPT_ONTIME * user_trips_pt
V_LATER = ASC_LATER + \
BETA_TT * trip_dur_later + \
BETA_WT * trip_walk_later + \
BETA_FREQ * trip_freq_later + \
BETA_IMPORTANT_LATER * user_imp_arr + \
BETA_GENDER_LATER * user_gender + \
BETA_AGE_LATER * AGE + \
BETA_INCOME_LATER * user_income + \
BETA_HOUSEHOLD_LATER * user_household + \
BETA_NPT_LATER * user_trips_pt
elif country == 'NL':
# Definition of utility functions - one for each alternative:
V_EARLIER = ASC_EARLIER + \
BETA_TT * trip_dur_earlier + \
BETA_WT * trip_walk_earlier + \
BETA_FARE_DISCOUNT * trip_discount_earlier + \
BETA_FREQ * trip_freq_earlier
V_ONTIME = ASC_ONTIME + \
BETA_TT * trip_dur_ontime + \
BETA_WT * trip_walk_ontime + \
BETA_FARE_DISCOUNT * trip_discount_ontime + \
BETA_FREQ * trip_freq_ontime + \
BETA_IMPORTANT_ONTIME * user_imp_arr + \
BETA_GENDER_ONTIME * user_gender
V_LATER = ASC_LATER + \
BETA_TT * trip_dur_later + \
BETA_WT * trip_walk_later + \
BETA_FARE_DISCOUNT * trip_discount_later + \
BETA_FREQ * trip_freq_later + \
BETA_IMPORTANT_LATER * user_imp_arr + \
BETA_GENDER_LATER * user_gender + \
BETA_SCOPE_LATER * trip_purpose
elif country == 'PT':
# Definition of utility functions - one for each alternative:
V_EARLIER = ASC_EARLIER + \
BETA_TT * trip_dur_earlier + \
BETA_WT * trip_walk_earlier + \
BETA_FREQ * trip_freq_earlier
V_ONTIME = ASC_ONTIME + \
BETA_TT * trip_dur_ontime + \
BETA_WT * trip_walk_ontime + \
BETA_FREQ * trip_freq_ontime + \
BETA_IMPORTANT_ONTIME * user_imp_arr + \
BETA_GENDER_ONTIME * user_gender + \
BETA_SCOPE_ONTIME * trip_purpose
V_LATER = ASC_LATER + \
BETA_TT * trip_dur_later + \
BETA_WT * trip_walk_later + \
BETA_FREQ * trip_freq_later + \
BETA_IMPORTANT_LATER * user_imp_arr + \
BETA_GENDER_LATER * user_gender + \
BETA_SCOPE_LATER * trip_purpose
else:
V_EARLIER = ASC_EARLIER
V_ONTIME = ASC_ONTIME
V_LATER = ASC_LATER
print('There is no model specification for ', country)
# Associate utility functions with the numbering of alternatives
V = {1: V_EARLIER, 2: V_ONTIME, 3: V_LATER}
# Associate the availability conditions with the alternatives. (Does not really apply on ToD modelling)
av = {
1: 1, # A user is able to arrive earlier..
2: 1, # A user is able to arrive on time..
3: 1
} # A user is able to arrive later..
# The choice model is a log logit, with availability conditions
logprob = bioLogLogit(util=V, av=av, choice=user_choice)
biogeme = bio.BIOGEME(database=estimationdb, formulas=logprob)
biogeme.modelName = "logitEstimation"
# Create the outputs of the estimation and store in a namedtuple (= Model)
results = biogeme.estimate()
betas = results.getBetaValues(
) # To be used later for the simulation of the model
structure = {
1: models.logit(V, av, 1),
2: models.logit(V, av, 2),
3: models.logit(V, av, 3)
}
Output = collections.namedtuple('Output',
['betas', 'structure', 'results'])
Model = Output(betas, structure, results)
self.__cleanup_after_model_training()
# print(self.evaluate_model(pandas_df_for_specified_country, Model))
return Model
