Automatic Word Sense Disambiguation Using Cooccurrence and Hierarchical Information

Fernández-Amorós, David; Heradio, Rubén; Somolinos, José Antonio Cerrada; Somolinos, Carlos Cerrada

doi:10.1007/978-3-642-13881-2_6

Cited by 6 publications

(3 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Co-occurrence information between words is also used in some WSD work [13,17,7,24,10,27]. Among them, Guthrie [13], Fernandez-Amoros [10] and Véronis [27] introduced unsupervised methods. Guthrie [13] proposed a two-level WSD on subject (domain) level and sense level (within a subject).…”

Section: Traditional Wsdmentioning

confidence: 99%

“…In each disambiguation level, they chose the subject/sense with the highest similarity between the context and the description of the subject/sense in a dictionary. Fernandez-Amoros [10] used cooccurrence to compute a relative matrix using mutual information (MI) measure. They used the MI between the target word and words of its context as the weight of each context words, and select the sense whose WordNet definition is most similar to the context by the bag-of-words model.…”

Section: Traditional Wsdmentioning

confidence: 99%

See 1 more Smart Citation

Wikification via link co-occurrence

Cai

Zhao

Zhu

et al. 2013

Proceedings of the 22nd ACM International Conference on Information &Amp; Knowledge Management

View full text Add to dashboard Cite

Wikification, which stands for the process of linking terms in a plain text document to Wikipedia articles which represent the correct meanings of the terms, can be thought of as a generalized Word Sense Disambiguation problem. It disambiguates multi-word expressions (MWEs) in addition to single words. Existing Wikification techniques either models the context of a given term as well as the Wikipedia article as bags of words, or compute global constraints among Wikipedia concepts by the link graph or link distributions. The first method doesn't achieve good results because the MWEs can have very different meanings than its constituent words which themselves are ambiguous. The second method doesn't produce high accuracy because the link structure or link distribution is often biased or incomplete by themselves due to the fact that Wikipedia pages are often sparsely linked. In this paper, we present a simple but powerful framework of sense disambiguation using cooccurrences of Wikipedia links in the Wikipedia corpus. We propose an iterative method to enrich the sparsely-linked articles by adding more links and then use the resulting link co-occurrence matrix to disambiguate an input document by a sliding window algorithm. Our prototype system achieves 89.97% precision and 76.43% recall on average for three benchmark data sets and compares favorably against all four state-of-the-art wikification techniques.1

show abstract

Section: Traditional Wsdmentioning

confidence: 99%

Section: Traditional Wsdmentioning

confidence: 99%

Wikification via link co-occurrence

Cai

Zhao

Zhu

et al. 2013

Proceedings of the 22nd ACM International Conference on Information &Amp; Knowledge Management

View full text Add to dashboard Cite

show abstract

“…The reason for this was that unrelated words rarely appear in both context and sense definition. A solution for this that involved extracting preference and layer information from the corpus to extend the sense annotation was proposed sometime later by Fernandez-Amoros et al [17].…”

Section: Related Workmentioning

confidence: 99%

Word sense disambiguation based on positional weighted context

Huang

Zheng

Kang

et al. 2012

Journal of Information Science

View full text Add to dashboard Cite

Word sense disambiguation (WSD) is a key factor in solving natural language processing problems. The purpose of WSD is to make computers automatically determine the specific meaning of a word in a specific context. In this regard, state-of-art studies have focussed on the co-occurrences of words to measure context similarity. However, a problem with these approaches is that they consider all the words within a certain range to have equal influence on the ambiguous word. In this paper, we propose a position-based algorithm for measuring context similarity. By assigning positional weights to context words, we compared the context similarity between a new instance and pre-labelled instances to determine the appropriate sense of the ambiguous word. Experiments on the Senseval-2 English lexical sample task showed that our algorithm can achieve good precision and recall. Even in a minimally supervised state, it performs well with few training instances.

show abstract