Using the Web to Obtain Frequencies for Unseen Bigrams

By Frank Keller, Mirella Lapata. International Conference on Computational Logic, 2003

402 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/5dfed29550d75cca99019aa52d40038dcb23b3cb

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Latent Variable Models of Selectional Preference (2010, methodology): “Table 2 presents results for a variety of predictive models – the Web frequencies reported by Keller and Lapata (2003) for two search engines, frequencies from the RASP-parsed BNC,4 the reimplemented methods of Rooth et al. (1999) and Padó et al. (2007), and the LDA, ROOTH-LDA and DUALLDA topic mode”
Random forests and the data sparseness problem in language modeling (2007, background): “Recent work by Keller and Lapata (2003) indicates that bigram frequencies from the web correlate reliably with bigram frequencies recreated by a class-based smoothing technique, at least for adjective–noun, noun–noun and verb–object bigrams.”
The Google Similarity Distance (2004, background): “Linguists judge the accuracy of Google counts trustworthy enough: In [23] (see also the many references to related research), it is shown that Web searches for rare two-word phrases correlated well with the frequency found in traditional corpora, as well as with human judgments of whether those phra”
VerbOcean: Mining the Web for Fine-Grained Semantic Verb Relations (2004, methodology): “Reliability of estimating bigram counts on the web via search engines has been investigated by Keller and Lapata (2003).”
Measuring semantic similarity between words using web search engines (2007, background): “Simple, unsupervised models demonstrably perform better when n - gram counts are obtained from the Web rather than from a large corpus [14, 15].”
Expectation-based syntactic comprehension. (2008, background): “…can be decomposed as follows (WS standing for Welches System):42 P (unterstützt|WS) = P (V.3sg|WS)P (unterstützt|V.3sg, WS) P (unterstützen|WS) = P (V.3pl|WS)P (unterstützen|V.3pl, WS) 40See Keller and Lapata (2003) for discussion of issues involved in obtaining n-gram frequencies from the Web.”