Vector Space Model

• Corpus forms a term-document matrix A

  • rows are terms
  • columns are documents

• Modern “large” corpus:

  • d = 107 documents (10Gb)
  • t = 105 terms

• Query also a term vector q

• Vector of query-document similarities: qTA

• Document-document similarities: ATA

• Improves significantly on Boolean

Previous slide

Next slide

Back to first slide

View graphic version