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

Table 3

Generative process for the labeled LDA model.

No.	Steps
1	For each topic k ∈ {1, ..., K}:
2	Generate Ψk = (ψk,1, ..., ψk,V)^T ∼ Dir(·|β)
3	For each driver d:
4	For each topic k ∈ {1, ..., K}
5	Generate Λ(m) k ∈ {0, 1} ∼ Bernoulli(·|Φk)
6	Generate
7	Generate θ^(m) = (,..., )^T ∼ Dir(|α^(m)
8	For each i in {1, ..., Nm}:
9	Generate zi ∈ {, ..., } ∼ Mult(·|θ^(m))
10	Generate  ∈ {1, ..., V} ∼ Mult(·|)