Conditional Random Fields - NER Notes

• NER - A named entity is a “real-world object” that’s assigned a name – for example, a person, a country, a product, or a book title. 
• CRF can take context into account.
• Each prediction is dependent only on its immediate neighbors. 
• CRF model to predict the conditional probability of Y by training the model parameters
• CRF builds transition probability that accounts for the likelihood of observing each transition between labels in the sequence
• CRF is a discriminative approach, It builds both likely transition and unlikely transitions
• A Discriminative model ‌models the decision boundary between the classes

Ref - Link

Feature Functions - Notes

Ref - Link

Ref - Link

Ref - Link

NER Approaches

Keep Exploring!!!