0
$\begingroup$

How many documents are generally required to produce good results for latent semantic indexing? By good results I mean relevant results are given for queries

$\endgroup$
2
$\begingroup$

This depends a lot on the specific application and the size of the vocabulary. It is reasonable to expect LSI to be good with 100,000 documents. But it can improve when you have millions of documents.

$\endgroup$
  • $\begingroup$ But is there a minimum threshold below which it absolutely wouldn't work. In terms of vocabulary, my application is a standard english vocabulary $\endgroup$ – user1893354 May 12 '16 at 16:06
  • $\begingroup$ There are no absolutes here. You will just get a gradual degradation in quality as your matrix becomes smaller. Only thing you can do is try it and see if it works and keep in mind that a typical application would have thousands of documents. $\endgroup$ – Aaron May 12 '16 at 18:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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