Unknown words such as proper nouns, abbreviations, and acronyms are a major obstacle in text processing. In particular, abbreviations are often used in specific domains. In this paper, we propose an automatic disabbreviation method using context information. In past research, a dictionary has conventionally been used to search abbreviation expansion candidates for an abbreviation. We use an abbreviation-poor text of the same domain instead of a dictionary. We calculate the plausibility of expansion candidates based on the similarity between the context of a target abbreviation and that of its expansion candidates. The similarity is calculated using the vector space model, in which each vector element consists of surrounding words. Experiments using about 10,000 documents in the aviation domain showed that the proposed method is superior to past methods by 10 in precision.
展开▼