An Open Access Journal
Table 1 Summary of possible traffic models and their identifiablity
From: Origin-destination matrix estimation with a conditionally binomial model
model | identifiability and comments |
|---|---|
Poisson, extended Poisson | identifiable, not compatible with Tampere data, used in [12] |
normal model (2-parameter models in general) | not identifiable; 6 parameters with 5 equations |
normal model with common variance | not identifiable; unable to separate local and through traffic |
conditionally Poissonian distributions, e.g., negative binomial distribution [14], Poisson-lognormal, Poisson - inverse Gaussian [18] | not identifiable; no bijection between the model parameters and the observation |
conditionally binomial model | identifiable, has similar properties to Tampere data |