LatentSemanticSearch: Perform a latent semantic analysis.

In joshua-heipel/tmbQuery: Search functions for Tambora.org

Description

This function compares the similarity between a low rank approximation of a term-document matrix and and a given query vector. (See e.g. "18 - Matrix decompositions and latent semantic indexing" in "Manning, S.D. et al. (2009): An Introduction to Information Retrieval. Cambridge University Press." for matematical details: https://nlp.stanford.edu/IR-book/pdf/irbookprint.pdf )

Usage

LatentSemanticSearch(tdm, qvec, spec = "nnn", dims = lsa::dimcalc_share(),
  ...)

Arguments

tdm,	term-document matrix as returned by the function createVSM()
qvec,	query vector as returned by the function createQuery()
spec,	determines weighting used by the function tm::weightSMART()
dims,	rank of the appoximation (integer value)

joshua-heipel/tmbQuery documentation built on May 8, 2019, 3:47 p.m.