3
$\begingroup$

I was watching a talk on latent semantic analysis and the speaker described experience applying LSA and REALLY messy data. He concluded that it demonstrated the difficulty of disambiguation of meaning, in regard to context. This got me thinking about how large the English vocabulary is.

Are other languages with smaller vocabularies, harder or easier to do LSA on?

$\endgroup$
1
  • 1
    $\begingroup$ English is not a romance language. $\endgroup$
    – Gala
    Jul 8, 2013 at 7:15

2 Answers 2

5
$\begingroup$

You have to keep in mind that the vocabulary size is not the only factor which will affect the LSA results.

In my experience a lot depends on the inflectional properties of a language. English is very LSA-friendly in this respect as it has virtually no inflection. Some Romance languages on the other hand still have pretty rich inflection so that having different strings for similar meanings will 'confuse' LSA unless you have a very good stemmer/lemmatizer.

If your primary concern is disambiguating the context you can try out some other techniques for discovering semantic structure. E.g. topic models (Latent Dirichlet Allocation; e.g. Blei et al., 2003) deal with homonymy much better.

If, theoretically, there were no other factors involved in addition to vocabulary size, I would expect that the ratio between vocabulary size and corpus size would matter more than vocabulary size alone.

$\endgroup$
3
$\begingroup$

Although I haven't tried solving word sense disambiguation with LSA, I do speak Hebrew and English. My guess is, that although Hebrew has a smaller vocabulary size than English, it would not be easier to perform algorithmic word sense disambiguation. The reason for this is that the size of the vocabulary is not as important as the linguistic patterns the language uses. For example, in Hebrew there are words that can be used to describe a concept and the exact opposite of that concept! I don't recall there are such words in English.

$\endgroup$
6
  • $\begingroup$ Does Hebrew have a smaller vocabulary than English? $\endgroup$ Jul 8, 2013 at 8:08
  • $\begingroup$ Biblical Hebrew vocab is ~1% the size of English vocab. $\endgroup$
    – Leeor
    Jul 8, 2013 at 8:16
  • $\begingroup$ Yes, there is Biblical Hebrew AND the modern, living language of Hebrew! When considering both, might the combined vocabulary be the same or greater than English? $\endgroup$ Jul 8, 2013 at 8:19
  • $\begingroup$ Modern day Hebrew is not large, it is a relatively young language and words are officially added by a small committee of experts. $\endgroup$
    – Leeor
    Jul 8, 2013 at 8:26
  • 2
    $\begingroup$ 'Cleave' (split or stick to) might be a candidate for an English word for both a concept & its opposite. 'Quite' too: compare 'quite intolerable' with 'quite warm'. I remember Quine discussed the matter, but I forget where. $\endgroup$ Jul 8, 2013 at 8:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.