Basic model

Trials

Find a preposition occurring in a sentence from some English text and write it down, together with the sentence context (e.g. <of,the price _ oil increased by five dollars>).

Note: requires an explicit definition of preposition occurrence, e.g. single word prepositions only (list?), not occurring as part of a multiword preposition, or as a verb particle or phrasal verb construction.

Sample space

The set of all pairs <x,y_z> where:

  • x is a listed preposition
  • yxz is a well-formed, meaningful sentence of English in which x functions as a preposition

Event space

The set of all subsets of the sample space.

For example:

  • <in,...> - the preposition chosen is "in"
  • <...,... price _ oil ...> - the preposition occurs immediately between words "price" and "oil"

Task

For each preposition-context pair <x,y_z> in the test set, verify that P(<x,...>|<...,y_z>) is greater than P(<x',...>|<...,y_z>), for all other prepositions x'.

Unigram model

Approximates P(<x,...>|<...,y_z>) as P(<x,...>).

Preceding bigram model

Approximates P(<x,...>|<...,y_z>) as P(<x,...>|<...,...w_...>) where w is a word.

--++ Following bigram model

Approximates P(<x,...>|<...,y_z>) as P(<x,...>|<...,..._w...>) where w is a word.

Surrounding trigram model

Approximates P(<x,...>|<...,y_z>) as P(<x,...>|<...,...w1_w2...>) where w1 and w2 are words.

-- MarkMcConville - 09 Sep 2008


This topic: TFlex > WebHome > Proposals > Prepositions > PrepositionNotes
Topic revision: r1 - 09 Sep 2008 - 19:41:26 - MarkMcConville
 
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback
This Wiki uses Cookies