Finite-State Sequence Tagging Model

A Finite-State Sequence Tagging Model is a Sequence Tagging Model (where each Atomic Element is Tagged) that is based on a Finite-State Automata.

• Context:
  • It can be used for:
    • a Text Segmentation Task.
  • It can be unable to Tag Overlapping Structures.
  • It can be unable to Tag Non-Contiguous Structures
• Example(s):
  • a Hidden Markov Model.
  • a Maximum Entropy Classification Model.
  • a ??Conditional Random Field.
• See: Structured Multilabel Classification Model.

---