915 use q0 and compute precision and recall graph

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

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Sec. 9.1.3 Viola*on of A2 Sec. 9.1.3   User does not have sufficient ini*al knowledge.   Examples:   Misspellings (Briiany Speers).   Cross ­language informa*on retrieval (hígado).   Mismatch of searcher’s vocabulary vs. collec*on vocabulary   Cosmonaut/astronaut Introduc)on to Informa)on Retrieval Relevance Feedback: Problems   There are several relevance prototypes.   Examples:   Long queries are inefficient for typical IR engine.   Long response *mes for user.   High cost for retrieval system.   Par*al solu*on:   Burma/Myanmar   Contradictory government policies   Pop stars that worked at Burger King ?   Users are oken reluctant to provide explicit feedback   It’s oken harder to understand why a par*cular document was retrieved aker applying relevance feedback   Report on contradictory government policies Evalua*on of relevance feedback strategies W hy   Only reweight certain prominent terms   Perhaps top 20 by term frequency   Oken: instances of a general concept   Good editorial content can address problem Introduc)on to Informa)on Retrieval Viola*on of A1 Relevance Feedback: Assump*ons Introduc)on to Informa)on Retrieval Introduc)on to Informa)on Retrieval Sec. 9.1.5   Use q0 and compute precision and recall graph   Use qm and compute precision recall graph   Assess on all documents in the collec*on   Spectacular improvements, but … it’s chea*ng!   Partly due to known relevant documents ranked higher   Must evaluate with respect to documents not seen by user   Use documents in residual coll...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online