|
Title:
|
EVALUATION OF AN INFORMATION RETRIEVAL MODEL BASED IN ANAPHORA RESOLUTION |
|
Author(s):
|
Francisco Santiago Do Carmo Pereira , Luana Vieira Morellato |
|
ISBN:
|
978-972-8924-93-5 |
|
Editors:
|
Pedro Isaías, Bebo White and Miguel Baptista Nunes |
|
Year:
|
2009 |
|
Edition:
|
2 |
|
Keywords:
|
Information Retrieval, Anaphora Resolution, Natural Language. |
|
Type:
|
Short Paper |
|
First Page:
|
334 |
|
Last Page:
|
338 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This article presents a structural information retrieval model based on the Discourse Nominal Structure (DNS). Text
representation used by this model allows a better text representation, once that the DNS theory can identify the main
entities presented by the author and the anaphoric bounds between them. Encoding the stored information in the Nominal
Structure the anaphora based information retrieval model can present in it's index only the most relevant terms showed by
the document. Classical Information Retrieval (IR) models considers only the document terms, which can lend to poor
retrieval because them does not considers the authors' anaphoric use or text's structure. This article intends to show the IR
quality that can be obtained once the IR model takes benefits of the discourse structure. It is made presenting an
experiment that compare a system based in anaphora based model with a system based in the Vector Space Model. |
|
|
|
|
|
|
