I am studying the different spatial modeling techniques that have so far used in safety analysis (modeling of road traffic accidents). Among the various statistical models used to model crash frequency, several studies utilized random effect models. After trying to understand the theory of such models and the difference between random effect and fixed effect models, some questions arised:
a) Some models include the term "Random effect: in their "names" (e.g. Random Effect Negative Binomial - RENB model), while some other do not despite the fact that include random effects/components (e.g. Bayesian Hierarchical Poisson Log-normal Models, with structured and an unstructured spatial random effects/component, temporal random effects/component and spatio-temporal random effects/component). Are all these models Random Effect models ?
b) Do Random Effect models account only for unobserved heterogeneity, only for spatial correlation or for both ?
c) Could Random Effect models be considered as spatial models?
d) Should I include Random Effect models in a review paper discussing spatial modeling techniques used in traffic safety analysis (modeling of crash frequencies)?
Thanks!!
If I might ask, have you considered Monte Carlo Simulation (I am most familiar with the Discrete Event Simulation implementation)? Upon running a given scenario multiple times, the output may be averaged or whatever makes the most sense. It's really effective when many parameters are involved.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS