Research Article

Driving Style Recognition under Connected Circumstance Using a Supervised Hierarchical Bayesian Model

Table 2

Mathematical notation for the labeled LDA model.

NotationDescription

ΘMultinomial distribution over driving styles
ΨMultinomial distribution over driving behaviors
ΑDirichlet prior parameters for Θ = (θ1, ..., θM)
ΒDirichlet prior parameters for Ψ = (ψ1, ..., ψK)
MNumber of drivers
NmNumber of driving behaviors in m-th drivers
KNumber of driving styles
The topic of the driving behavior
The ith driving behavior
Λ(m)The label vector of driver m
ΦkLabel prior for topic k