Text Segment Predictor Feature

A Text Segment Predictor Feature is a Sequence Segmentation Predictor Feature for text segments.

• AKA: Text Segmentation Feature.
• Context:
  • It can be:
    • a Text Token-based Predictor Feature.
    • a Text Window-based Predictor Feature.
• Example(s):
  • the whole entity has an exact match in an database of entities.
  • the length of the entity is between 4 and 8 words.
  • the third or fourth token of the entity is a “-”
  • the last three tokens in the entity are digits.
  • …
• Counter-Example(s):
  • Text Tagging Predictor Feature.
• See: Tagging Feature.

---

---

References

2006

• (Sarawagi, 2006) ⇒ Sunita Sarawagi. (2006). “Efficient Inference on Sequence Segmentation Models.” In: Proceedings of the 23rd International Conference on Machine Learning (ICML 2006). doi:10.1145/1143844.1143944
  • … Examples of such segment level features for extraction tasks are: the whole entity has an exact match in an database of entities, the length of the entity is between 4 and 8 words, the third or fourth token of the entity is a “-”, and the last three tokens in the entity are digits. Sequence segmentation models provide a direct and natural way of encapsulating all such entity features.

---