IR-part2

# 3 0 0 1 1 1 052 0 best 1 1 50000 13 13 034 0 0 0 0 0

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: maximal similarity. §༊  Key idea: Rank documents according to angle with query. Introduc)on to Informa)on Retrieval Sec. 6.3 From angles to cosines §༊  The following two no*ons are equivalent. §༊  Rank documents in decreasing order of the angle between query and document §༊  Rank documents in increasing order of cosine(query,document) §༊  Cosine is a monotonically decreasing func*on for the interval [0o, 180o] Introduc)on to Informa)on Retrieval Sec. 6.3 From angles to cosines §༊  But how – and why – should we be compu*ng cosines? Introduc)on to Informa)on Retrieval Sec. 6.3 Length normaliza*on §༊  A vector can be (length- ) normalized by dividing each of its components by its length – for this we use the L2 norm: 2 x2= ∑x ii §༊  Dividing a vector by its L2 norm makes it a unit (length) vector (on surface of unit hypersphere) §༊  Eﬀect on the two documents d and dʹȃ (d appended to itself) from earlier slide: they have iden*c...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online