Post on 04-Jan-2016
transcript
Critical Issues in Estimating and Applying Nested Logit Mode
Choice Models
Ramachandran BalakrishnaSrinivasan Sundaram
Caliper Corporation
12th TRB National Transportation Planning Applications Conference, Houston, Texas
19th May, 2009
Outline
• Introduction• Motivation• Non-uniqueness in model
estimation• Choice of utility scaling method• Numerical example• Conclusion• References
Introduction
• Nested Logit (NL): popular for mode choice
• Captures unobserved shared effects across modes
• Requires estimation from disaggregate data− Unknowns:
− Utility coefficients, nest thetas− Software:
− Biogeme, ALOGIT, TransCAD, etc.
Motivation: Highlight critical NL issues• Estimates may not be unique− Coefficients unique only for fixed thetas
• Daganzo and Kusnic (1992)
− Final estimates depend on starting thetas• Koppelman & Bhat (2006)• Wide range of estimates possible
• Utilities must be ‘scaled’− Parent thetas are built into the utilities− Utilities need scaling before comparing across
nests− Estimation programs use different scaling
methods− Some are inconsistent with utility
maximization− Koppelman & Wen (1998)
Non-Uniqueness in Model Estimation (I)• Example (from Koppelman & Bhat, 2006)− Three sets of starting theta values
• Results very sensitive to starting theta values
• Very similar or identical final LL likely− Harder to select a ‘good’ model• Unrealistic estimates possible
Starting thetas Non-motorized 0.5 0.2 1.0 Auto 0.5 0.7 1.0 Final thetas
Non-motorized Auto
0.5799.00E-05
0.2260.703
0.2270.703
Constants Transit -1.34 -0.17 -0.169 Shared Ride 2 -0.0001 -0.282 -0.282 ln(Persons per HH) Transit 0.545 0.899 0.9 Shared Ride 2 0.0001 0.266 0.267 Travel time Motorized -9.00E-07 -0.0246 -0.0246 Non-motorized -0.0762 -0.08 -0.08 Final log-likelihood (LL) -4450.57 -4447.48 -4447.48
Non-Uniqueness in Model Estimation (II)• Model selection checks and guidelines
– Final log-likelihood need not be only criterion• Coefficient magnitudes, signs• Relevant ratios (e.g. value of time)• Elasticities (within and across nests)
• Must re-estimate with various starting thetas
−Pick the best possible model−Detailed multi-dimensional search
• One option: grid search• Implemented in TransCAD 5.0
Utility Scaling
• Basic NL formulation
– effects built into utilities– Difficult to compare utilities across nests– Counter-intuitive direct, cross elasticities– Inconsistent with utility maximization
– Solution: scale utilities to remove effects• Two scaling approaches
)ln(
)| (
CPDA
CPDA
DA
VV
VV
V
eeLogsumAuto
ee
eAutoAloneDriveP
Utility Scaling Methods (I)
• Scale by parent
− Consistent with utility maximization− Intuitive direct and cross elasticities− Implemented in TransCAD 5.0
AutoCPAutoDA
AutoCPAutoDA
AutoDA
VV
VV
V
eeLogsumAuto
ee
eAutoAloneDriveP
//
//
/
ln
)| (
Utility Scaling Methods (II)
• Scale by product of ’s− Requires dummy nests, constraints on ‘s− Harder to apply and interpret− ALOGIT
Auto
VV
VV
V
CPDA
CPDA
DA
eeLogsumAuto
ee
eAutoAloneDriveP
Motorized
//
//
/
where
ln
)| (
Utility Scaling Methods (III)
• Choice of scaling method impacts mode shares− Identical only for models with two levels of nests
• Estimation– Utility maximization requires scaling by parent
• Model application– Critical to know how model was estimated!
• TransCAD 5.0– Estimation options: no scaling, scale by parent – Application options: all three methods
Numerical Example (I)
• TransCAD 5.0 (Caliper Corporation, 2008)
− Estimates and applies NL, MNL models− Batch-enabled for efficient theta search− Estimates select coefficients while fixing
others− Allows different scaling methods
− Has intuitive GUI
• Automatically combines different data sources− Surveys, zonal tables, matrices, etc.
• Efficiently handles market segments
Numerical Example (II)
• Travel survey (Southern California Assoc. of Govts., SCAG)
• 9885 survey records (home-based work trips)• Modes: Non-Motorized, Drive Alone, Carpool,
Transit• Utilities scaled by parent • 101 estimations of starting in [0,1] , 0.01
step size• 52 valid runs with final in [0,1]
− Almost identical log-likelihood,
Numerical Example (III)Results: Constants for DA, CP, NM
Numerical Example (IV)
Results: Coefficients of No_License Dummy, Walk Time
Numerical Example (V)
Results: Estimated theta values
Theta_Auto (initial)
Theta_Auto (estimated)
0.01 0.012070.02 -8.844380.59 0.0121230.6 -8.910630.99 0.012115
0.9999 -8.84436
Conclusion
• More care is required in estimating and applying Nested Logit mode choice models
• Good practice is to perform extensive estimation runs
• One should match the scaling used in estimation and application
References
Caliper Corporation (2008) Travel Demand Modeling with TransCAD, Version 5, Newton, MA.
C. F. Daganzo and M. Kusnic (1992) Another Look at the Nested Logit Model, UC Berkeley report UCB-ITS-RR-92-2.
F. S. Koppelman and C. Bhat (2006) A Self Instructing Course in Mode Choice Modeling: Multinomial and Nested Logit Models, prepared for U.S. DOT, FTA.
F. S. Koppelman and C-H. Wen (1998) Alternative Nested Logit Models: Structure, Properties and Estimation. Transportation Research 32B, No. 5, pp. 289-298.