Bayesian Disaggregate and Aggregate Calibration of Path Logit Choice Models

Table 1

Calibration results for the path choice model, disaggregate data.

ID, approach	1, C	2, C	3, B	4, B
Interviews	124	186	124	186
θ T ∈ ]0, +∞[(1/h): travel time (t-Student)	−0.55 (−1.53)	−0.45 (−1.46)	−0.77 (−2.15)	−0.64 (−2.09)
θ P ∈ [0, 1]: percentage on motorway (t-Student)	1.57 (3.38)	1.78 (4.54)	1.67 (3.89)	1.78 (4.93)
θ L ∈ {0, 1}: label time (t-Student)	10.98 (0.15)	14.19 (0.06)	2.61 (3.26)	2.96 (4.06)
ln (likelihood)^∧	−71.71	−106.83	−72.55	−108.01
Pseudo-ρ^2	0.388	0.394	0.381	0.388

C, Classic; B, Bayesian; ^statistics with parameters equal to zero; ln (likelihood): −117.17 with 124 interviews; 176.36 with 186 interviews.