MIT Press

Table 2

Feature templates used for the Bio-NER task. w_i is the current word token on position i. t_i is the POS tag on position i. o_i is the orthography mode on position i. y_i is the classification label on position i. y_{i − 1}y_i represents label transition. represents a Cartesian product between two sets.

Word Token–based Features:

{w_{i − 2}, w_{i − 1}, w_i, w_{i + 1}, w_{i + 2}, w_{i − 1}w_i, w_iw_{i + 1}}

×{y_i, y_{i − 1}y_i}

Part-of-Speech (POS)–based Features:

{t_{i − 2}, t_{i − 1}, t_i, t_{i + 1}, t_{i + 2}, t_{i − 2}t_{i − 1}, t_{i − 1}t_i, t_it_{i + 1}, t_{i + 1}t_{i + 2}, t_{i − 2}t_{i − 1}t_i, t_{i − 1}t_it_{i + 1}, t_it_{i + 1}t_{i + 2}}

×{y_i, y_{i − 1}y_i}

Orthography Pattern–based Features:

{o_{i − 2}, o_{i − 1}, o_i, o_{i + 1}, o_{i + 2}, o_{i − 2}o_{i − 1}, o_{i − 1}o_i, o_io_{i + 1}, o_{i + 1}o_{i + 2}}

×{y_i, y_{i − 1}y_i}

View Large

This Feature Is Available To Subscribers Only