The NLP task at INEX 2004

Authors:
Shlomo Geva;Tony Sahama
Affiliations:
Queensland University of Technology, Brisbane, Australia;Queensland University of Technology, Brisbane, Australia
Venue:
ACM SIGIR Forum
Year:
2005

Citing 0
Cited 1

Web document modeling

The adaptive web

Quantified Score

Hi-index	0.00

Visualization

Abstract

The INEX workshop is concerned with Evaluating the effectiveness of XML retrieval systems. In 2004 a natural language query task was added to the INEX Ad hoc track. Standard INEX Ad hoc topic titles are specified in NEXI -- a simplified and restricted subset of XPath, with a similar feel, and yet with a distinct IR flavour and interpretation. The syntax of NEXI is rigid and it imposes some limitations on the kind of information need that it can faithfully capture. At INEX 2004 the NLP question to be answered was simple -- is it practical to use a natural language query that is the equivalent of the formal NEXI title? The results of this experiment are reported and some information on the future direction of the NLP task is presented.