Download Citation on ResearchGate | Automatic Acquisition of Hyponyms from Large Text Corpora | We describe a method for the automatic. Automatic Acquisition of Hyponyms from Large Text Corpora. Anthology: C ; Volume: COLING Volume 2: The 15th International Conference on. This post is a review of the paper: Hearst, Marti A. “Automatic acquisition of hyponyms from large text corpora. In Proceedings of the.
|Published (Last):||2 February 2007|
|PDF File Size:||11.63 Mb|
|ePub File Size:||20.1 Mb|
|Price:||Free* [*Free Regsitration Required]|
When comparing against WordNet, three outcomes were considered.
Topics Discussed in This Paper.
Automatic Acquisition of Hyponyms from Large Text Corpora
Grolier Electronic Publishing, Danbury…. Notify me of new comments via email. You are commenting using your Automayic account.
When comparing to WordNet, relations were restricted to only nouns without modifiers.
Fill in your details below or click an icon to log in: Patterns The approach is based on pattern matching. One reason was due the type of data contained in WordNet, but it also was suggested in general that it is difficult to know which modifiers are important to the relation. This information may have been contained in a previous sentence.
For example, the was found where steatornis is a species of bird. This paper looks at extracting information from raw text. You are commenting using your WordPress. Sorry, your blog cannot share posts by atuomatic.
Citation Statistics 3, Citations 0 ’91 ’97 ’04 ’11 ‘ Two goals motivate the approach: Shortcomings When comparing to WordNet, relations were restricted to only nouns without modifiers. BrentRobert C. Appositives were difficult to match accurately.
From This Paper Figures, tables, and topics from this paper.
Automatic Acquisition of Hyponyms from Large Text Corpora – Semantic Scholar
WordNet contains 34, noun forms and 26, synsets. Reconciling information contained in separate sentences may be challenging with pattern recognition alone.
See our FAQ for additional information. To find out more, including how to control cookies, see here: Text corpus Search for additional papers on this topic. References Publications referenced by this paper. Find locations in the text corpus where these expressions occur near each other. A common issue was underspecification. It does not require acqhisition nor context specific, preencoded knowledge.
If both words were in WordNet but the relation was not, then a new hyponym connection was suggested. Email required Address never made public. Lastly, if one or both noun phrases were not in WordNet, then the words and their relation were suggested. Semantic Scholar estimates that this publication has 3, citations based on the available data.
We identify a set of lexicosyntactic patterns that are easily recognizable, that occur frequently and across text genre boundaries, and that indisputably indicate the lexical relation of interest.
The approach described in this paper is different in that only one sample of a relation needs to be found in a text to be useful. Statistical approaches have ccorpora been used that look to determine lexical relations by looking at very large text samples.
Other types of relations were tried without success.