Smooth
Smoothing document length  
Home
People
Latest Progress
Publications
FAQ
Search our site:
 
powered by Terrier and
Labrador

department of computing science at the university of glasgow
university of glasgow

The Smooth Project

The Smooth project aims to optimise the performance of an Information Retrieval (IR) system by tuning the free parameters of the term frequency normalisation methods.

The setting of the term frequency normalisation parameter suffers from the query-dependent and collection-dependent problems, which significantly hurt the robustness of the retrieval performance. In this project, we proposed to provide a theoretical understanding of the term frequency normalisation, as well as solutions for the two problems.

Smooth is funded by the Leverhulme Trust, number F/00179/S.

Links
Information Retrieval Group
Labrador
Terrier
The Leverhulme Trust Terrier Logo

Last Update: 9 June, 2005