** Next:** When does relevance feedback
** Up:** Relevance feedback and pseudo
** Previous:** The #rocchio71### algorithm.
** Contents**
** Index**

##

Probabilistic relevance feedback

Rather than reweighting the query in a vector space, if a user has told
us some relevant and nonrelevant documents, then we can proceed to
build a . One way of doing this is with a
Naive Bayes probabilistic model.
If is a Boolean indicator variable expressing
the relevance of a document, then we can estimate , the probability of a
term appearing in a document, depending on whether it is relevant
or not, as:

where is the total number of documents, is the number that
contain , is the set of known relevant documents, and is the subset of this set containing . Even though the set of known relevant documents is a perhaps small subset of the true set of relevant documents, if we assume that the set of relevant documents is a small subset of the set of all documents then the estimates given above will be reasonable.
This gives a basis for another way of changing the query term
weights. We will discuss such probabilistic approaches more in
Chapters 11 13 , and in particular outline the application to relevance feedback in Section 11.3.4 (page ). For the moment, observe that using just Equation 50 as a basis for term-weighting is likely insufficient.
The equations use only collection statistics and information about the term
distribution within the
documents judged relevant. They preserve no memory of the original query.

** Next:** When does relevance feedback
** Up:** Relevance feedback and pseudo
** Previous:** The #rocchio71### algorithm.
** Contents**
** Index**
© 2008 Cambridge University Press

This is an automatically generated page. In case of formatting errors you may want to look at the PDF edition of the book.

2009-04-07