Named Entity Recognition using an HMM-based Chunk Tagger

By Guodong Zhou, Jian Su. Annual Meeting of the Association for Computational Linguistics, 2002

805 citations recorded by the source corpus.

This paper proposes a Hidden Markov Model (HMM) and an HMM-based chunk tagger, from which a named entity (NE) recognition (NER) system is built to recognize and classify names, times and numerical quantities. Through the HMM, our system is able to apply and integrate four types of internal and external evidences: 1) simple deterministic internal feature of the words, such as capitalization and digitalization; 2) internal semantic feature of important triggers; 3) internal gazetteer feature; 4) external macro context feature. In this way, the NER problem can be resolved effectively. Evaluation of our system on MUC-6 and MUC-7 English NE tasks achieves F-measures of 96.6% and 94.1% respectively. It shows that the performance is significantly better than reported by any other machine-learning system. Moreover, the performance is even consistently better than those based on handcrafted rules.

Subfields: Artificial Intelligence, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/51bdd0b839fe8ab1daafa05d1ed03edff601f4c7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Management Science and Operations Research

Citations in the corpus

A Survey on Deep Learning for Named Entity Recognition (2018, methodology): “Word-level features (e.g., case, morphology, and part-of-speech tag) [56]–[58], list lookup features (e.g., Wikipedia gazetteer and DBpedia gazetteer) [59]–[62], and document and corpus features (e.g., local syntax and multiple occurrences) [63]–[66] have been widely used in various supervised NER…”
TwiNER: named entity recognition in targeted twitter stream (2012, background): “In most of the cases, they depend on the Partof-Speech (POS) tags, which again need a tagger to be trained with supervised approach based on linguistic features[14, 20, 21].”
Learning to Automatically Solve Logic Grid Puzzles (2015, methodology): “research shares many similarity with Named Entity Recognition (Nadeau and Sekine, 2007; Zhou and Su, 2002) and the Word Sense disambiguation (Stevenson and Wilks, 2003; Sanderson, 1994) task.”
Interacting meaningfully with machine learning systems: Three experiments (2009, methodology): “We coded feedback in this category if the necessary knowledge could be learned from analysis of large document collections or obtained from other online resources (e.g., Wordnet (Miller 1995) or named-entity recognizers (Zhou and Su 2002)).”
Adaptive Co-attention Network for Named Entity Recognition in Tweets (2018, methodology): “Zhou and Su (2002) proposed the use of an HMM-based chunk tagger.”
Improving Multimodal Named Entity Recognition via Entity Span Detection with Unified Multimodal Transformer (2020, methodology): “Although many methods coupled with either discrete shallow features (Zhou and Su, 2002; Finkel et al., 2005; Torisawa et al., 2007) or continuous deep features (Lample et al., 2016; Ma and Hovy, (a).”
Named Entity Recognition with a Maximum Entropy Approach (2003, background): “Name Class of Previous Occurrences The name class of previous occurrences of is used as a feature, similar to (Zhou and Su, 2002).”
Unsupervised biomedical named entity recognition: Experiments with clinical and biological texts (2013, methodology): “With sequence labeling models like Hidden Markov Models (HMMs) and Conditional Random Fields (CRFs), the two tasks could be jointly handled taking advantage of the Markov property which models transitions between labels [15,16].”
Semi-Supervised Learning for Natural Language (2005, positive): “Generative models such as Hidden Markov Models (Bikel et al., 1999; Zhou and Su, 2002) have shown excellent performance on the Message Understanding Conference (MUC) datasets (Sundheim, 1995; Chinchor, 1998).”
Toward harnessing user feedback for machine learning (2007, neutral): “, Wordnet [17] or named-entity recognizers [27]).”
A Comprehensive Survey on Automatic Knowledge Graph Construction (2023, background): “Hence, applications built on hidden Markov models [93] and conditional random fields (CRF) models [94], which capture neighborhood-dependencies, are popular NER designs.”
A Survey on Recent Advances in Named Entity Recognition from Deep Learning models (2018, methodology): “Zhou and Su (2002) used HMM (Rabiner and Juang, 1986; Bikel et al., 1997) an NER system on MUC-6 and MUC-7 data, achieving 96.6% and 94.1% F score, respectively.”