BibTeX
@MISC{H_semanticextensions,
author = {Ben H and Rohini Rajaraman},
title = {Semantic Extensions to Syntactic Analysis of Queries},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We intend to show that leveraging semantic features can improve precision and recall of query results in information retrieval (IR) systems. Nearly all existing IR systems are built around the 'bag-of-words ' model. One major shortcoming of this model is that semantic information in documents and queries is completely ignored. It has been shown in a limited setting that some syntactic information can be useful to IR [1]. We intend to show that leveraging semantic features such as synonyms can improve precision and recall of query results, especially when used in conjunction with syntactic relationships. In this paper, we describe the various machine learning algorithms and feature sets that we use to learn a binary relevance classification function. At this stage, we are not attempting to build an actual IR system using our techniques, but we hope to show the benefits of such a system.
