Ranking with uncertain scoring functions
Davide Martinenghi (Politecnico di Milano)
Info
|
Date |
17th May 2011 (week , Trinity Term 2011) |
|
Time |
12:00 |
|
Place |
380 |
Abstract
The goal of ranking queries is to report the top K results according to a user-defined scoring function. Commonly, scoring
functions are expressed as a weighted summation of multiple scores. However, users are often unable to precisely specify
the exact weights in such functions. Yet, they can easily provide approximations of the weights or identify their relative
importance. For this reason, uncertain/incomplete scoring functions suitably capture user preferences in this scenario.
In this talk, I will discuss the semantics of ranking queries with uncertain scoring functions. In addition, I will address
the problem of managing the sensitivity of computed results with respect to refinements made by the user.
Further info
|
Related series |
|
