|
Title:
|
AN APPROACH TO PREVENT STEMMING SIDE EFFECTS IN INFORMATION RETRIEVAL |
|
Author(s):
|
Ahmet Arslan, Ozgur Yilmazel |
|
ISBN:
|
978-972-8939-30-4 |
|
Editors:
|
Hans Weghorn, Pedro Isaías and Radu Vasiu |
|
Year:
|
2010 |
|
Edition:
|
Single |
|
Keywords:
|
Lucene, precision, porter, recall, stemming. |
|
Type:
|
Reflection Paper |
|
First Page:
|
275 |
|
Last Page:
|
277 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Traditional usage of stemming in Information Retrieval increases recall at the expense of harming precision. However in Web Search scenario precision is more important. In this paper, we propose an approach that applies stemming with aim to improve search recall without significant loss in precision. Our proposed solution keeps original term along with its stem and gives smaller weight to stem. Implementation of our approach is done in Lucene which is an open source full text search library. Our experiments on TREC4 ad hoc task environment show that stemming hurt precision for 16 topics and we can improve precision by 26% on these topics and 7.8% for over all topic set. |
|
|
|
|
|
|
