A Context Pattern Induction Method for Named Entity Extraction

Proceedings of the 10th Conference on Computational Natural Language Learning (CoNLL-X), pages 141–148, New York City, June 2006. c⃝2006 Association for Computational Linguistics A Context Pattern Induction Method for Named Entity Extraction Partha Pratim Talukdar CIS Department University of Pennsylvania Philadelphia, PA 19104 partha@cis.upenn.edu Thorsten Brants Google, Inc. 1600 Amphitheatre Pkwy. Mountain View, CA 94043 brants@google.com Mark Liberman Fernando Pereira CIS Department University of Pennsylvania Philadelphia, PA 19104 {myl,pereira}@cis.upenn.edu Abstract We present a novel context pattern in- duction method for information extrac- tion, specifically named entity extraction. Using this method, we extended several classes of seed entity lists into much larger high-precision lists. Using token member- ship in these extended lists as additional features, we improved the accuracy of a conditional random field-based named en- tity tagger. In contrast, features derived from the seed lists decreased extractor ac- curacy. 1 Introduction Partial entity lists and massive amounts of unla- beled data are becoming available with the growth of the Web as well as the increased availability of specialized corpora and entity lists. For example, the primary public resource for biomedical research, MEDLINE, contains over 13 million entries and is growing at an accelerating rate. Combined with these large corpora, the recent availability of entity lists in those domains has opened up interesting op- portunities and challenges. Such lists are never com- plete and suffer from sampling biases, but we would like to exploit them, in combination with large un- labeled corpora, to speed up the creation of infor- mation extraction systems for different domains and languages. In this paper, we concentrate on explor- ing utility of such resources for named entity extrac- tion. Currently available entity lists contain a small fraction of named entities, but there are orders of magnitude more present in the unlabeled data1. In this paper, we test the following hypotheses: i. Starting with a few seed entities, it is possible to induce high-precision context patterns by ex- ploiting entity context redundancy. ii. New entity instances of the same category can be extracted from unlabeled data with the in- duced patterns to create high-precision exten- sions of the seed lists. iii. Features derived from token membership in the extended lists improve the accuracy of learned named-entity taggers. Previous approaches to context pattern induc- tion were described by Riloff and Jones (1999), Agichtein and Gravano (2000), Thelen and Riloff (2002), Lin et al. (2003), and Etzioni et al. (2005), among others. The main advance in the present method is the combination of grammatical induction and statistical techniques to create high-precision patterns. The paper is organized as follows. Section 2 de- scribes our pattern induction algorithm. Section 3 shows how to extend seed sets with entities extracted by the patterns from unlabeled data. Section 4 gives experimental results, and Section 5 compares our method with previous work. 1For example, based on approximate matching, there is an overlap of only 22 organizations between the 2403 organiza- tions present in CoNLL-2003 shared task training data and the Fortune-500 list. 141 2 …

Content truncated to 500 words