Domain adaptation with latent semantic association for named entity recognition

Abstract

Domain adaptation is an important problem in named entity recognition (NER). NER classifiers usually lose accuracy in the domain transfer due to the different data distribution between the source and the target domains. The major reason for performance degrading is that each entity type often has lots of domainspecific term representations in the different domains. The existing approaches usually need an amount of labeled target domain data for tuning the original model. However, it is a labor-intensive and time-consuming task to build annotated training data set for every target domain. We present a domain adaptation method with latent semantic association (LaSA). This method effectively overcomes the data distribution difference without leveraging any labeled target domain data. LaSA model is constructed to capture latent semantic association among words from the unlabeled corpus. It groups words into a set of concepts according to the related context snippets. In the domain transfer, the original term spaces of both domains are projected to a concept space using LaSA model at first, then the original NER model is tuned based on the semantic association features. Experimental results on English and Chinese corpus show that LaSA-based domain adaptation significantly enhances the performance of NER.

Cross-Domain Sentiment Classification Using a Sentiment Sensitive Thesaurus

“…Domain adaptation in general has been studied in various other tasks such as part of speech tagging [39], named entity recognition [40], noun phrase chunking [34] and dependency parsing [41]. Domain adaptation methods can be broadly classified into fully supervised and semi-supervised adaptation [34].…”

Multi-Label Classification of Contributing Causal Factors in Self-Reported Safety Narratives

“…Effective classification of textually based data remains an active area of research in both computer science and the study of safety taxonomies. TheF 1 scores here are slightly lower than other studies that combined the use of LSA and nearest neighbor labeling [19,20]. In a study of named entity recognition (identification of person, location, and organization), Guo et al [19] returned F 1 scores between 0.50 and 0.70, dependent on the domain analyzed.…”

A Multi-Domain Web Text Feature Extraction Model for e-Science Environment

“…The model with Web gazetteer, using a relatively simple way to extract geographic time information, combined with space time, to get a more comprehensive description of geographic information metadata. Honglei Guo and Huijia Zhu [7] (2009) present a named entity detection model based on semantic association analysis. The model effectively overcame the differences of the distribution of data among different domains by mining the latent semantic association between words, contributed to the accuracy of entity identification.…”