GenPark Hidden Markov Model Viterbi Decoder Skill

Viterbi dynamic programming algorithm for decoding maximum a posteriori (MAP) state sequences in Hidden Markov Models.

Check out GenPark and the GenPark MCP Catalog.

graph LR
    subgraph t=0
        H0[HEALTHY]
        F0[FEVER]
    end
    subgraph t=1
        H1[HEALTHY]
        F1[FEVER]
    end
    subgraph t=2
        H2[HEALTHY]
        F2[FEVER]
    end
    H0 -->|Viterbi Trellis| H1
    H1 -->|Viterbi Trellis| F2
    style H0 fill:#e8f5e9
    style H1 fill:#e8f5e9
    style F2 fill:#ffebee

Loading

Features

• Dynamic programming log-space calculations avoiding underflow.
• Exact backpointer sequence path reconstruction.
• Pure Python standard library.