A Bootstrapping Approach to Named Entity Classification Using Successive Learners

A Bootstrapping Approach to Named Entity Classification Using Successive Learners Cheng Niu, Wei Li, Jihong Ding, Rohini K. Srihari Cymfony Inc. 600 Essjay Road, Williamsville, NY 14221. USA. {cniu, wei, jding, rohini}@cymfony.com Abstract This paper presents a new bootstrapping approach to named entity (NE) classification. This approach only requires a few common noun/pronoun seeds that correspond to the concept for the target NE type, e.g. he/she/man/woman for PERSON NE. The entire bootstrapping procedure is implemented as training two successive learners: (i) a decision list is used to learn the parsing-based high precision NE rules; (ii) a Hidden Markov Model is then trained to learn string sequence-based NE patterns. The second learner uses the training corpus automatically tagged by the first learner. The resulting NE system approaches supervised NE performance for some NE types. The system also demonstrates intuitive support for tagging user-defined NE types. The differences of this approach from the co-training-based NE bootstrapping are also discussed. 1 Introduction Named Entity (NE) tagging is a fundamental task for natural language processing and information extraction. An NE tagger recognizes and classifies text chunks that represent various proper names, time, or numerical expressions. Seven types of named entities are defined in the Message Understanding Conference (MUC) standards, namely, PERSON (PER), ORGANIZATION (ORG), LOCATION (LOC), TIME, DATE, MONEY, and PERCENT1 (MUC-7 1998). 1 This paper only focuses on classifying proper names. Time and numerical NEs are not yet explored using this method. There is considerable research on NE tagging using different techniques. These include systems based on handcrafted rules (Krupka 1998), as well as systems using supervised machine learning, such as the Hidden Markov Model (HMM) (Bikel 1997) and the Maximum Entropy Model (Borthwick 1998). The state-of-the-art rule-based systems and supervised learning systems can reach near-human performance for NE tagging in a targeted domain. However, both approaches face a serious knowledge bottleneck, making rapid domain porting difficult. Such systems cannot effectively support user-defined named entities. That is the motivation for using unsupervised or weakly- supervised machine learning that only requires a raw corpus from a given domain for this NE research. (Cucchiarelli & Velardi 2001) discussed boosting the performance of an existing NE tagger by unsupervised learning based on parsing structures. (Cucerzan & Yarowsky 1999), (Collins & Singer 1999) and (Kim 2002) presented various techniques using co-training schemes for NE extraction seeded by a small list of proper names or handcrafted NE rules. NE tagging has two tasks: (i) NE chunking; (ii) NE classification. Parsing- supported NE bootstrapping systems including ours only focus on NE classification, assuming NE chunks have been constructed by the parser. The key idea of co-training is the separation of features into several orthogonal views. In case of NE classification, usually one view uses the context evidence and the other relies on the lexicon evidence. Learners corresponding to different views learn from each other iteratively. One issue of co-training is the error propagation problem in the process of the iterative learning. The rule precision drops iteration-by-iteration. In the early …

Content truncated to 500 words