Named Entity Extraction with Finite State Transducers
This is an incremental improvement for NLP tasks, offering a simple, language-agnostic approach to named entity extraction.
The paper tackles named entity tagging by constructing a series of automatons using supervised learning, achieving an F1 score of 60% on the Spanish CoNLL-2002 dataset.
We describe a named entity tagging system that requires minimal linguistic knowledge and can be applied to more target languages without substantial changes. The system is based on the ideas of the Brill's tagger which makes it really simple. Using supervised machine learning, we construct a series of automatons (or transducers) in order to tag a given text. The final model is composed entirely of automatons and it requires a lineal time for tagging. It was tested with the Spanish data set provided in the CoNLL-$2002$ attaining an overall $F_{β= 1}$ measure of $60\%.$ Also, we present an algorithm for the construction of the final transducer used to encode all the learned contextual rules.