Login

Publications  •  Project Statistics

Glossary  •  Schools  •  Disciplines
People Search: 
   
Title/Abstract Search: 

Dissertation Information for Jane Greenberg

NAME:
- Jane Greenberg

DEGREE:
- Ph.D.

DISCIPLINE:
- Library and Information Science

SCHOOL:
- University of Pittsburgh (USA) (1998)

ADVISORS:
- Edie Rasmussen

COMMITTEE MEMBERS:
- Louise Su
- Sara Fine
- Lei Zeng

MPACT Status: Fully Complete

Title: An examination of the impact of lexical-semantic relationships on retrieval effectiveness during the query expansion (QE) process

Abstract: This study examined how vocabulary terms, generated by semantic relationships in a structured thesaurus, influenced retrieval effectiveness when used to automatically expand a query in an operational environment. This study also compared the impact of semantic relationships on retrieval effectiveness for automatic query expansion (AQE) and interactive query expansion (IQE) processes.

The test environment included the business periodical database ABI / Inform , the ProQuest Controlled Vocabulary , which is the thesaurus used to index and retrieve ABI / Inform documents, and a group of M.B.A. students at the University of Pittsburgh, who had queries they wanted to pursue in a business periodical database.

Significant results were found for AQE via certain semantic relationships and for the AQE and IQE comparison. (1) AQE via all the semantic relationships increased relative recall. (2) AQE via related terms (RTs) and broader terms (BTs) augmented relative recall more than synonyms (SYs) and narrower terms (NTs). (3) SYs and NTs achieved better precision results than RTs and BTs. The loss in precision for AQE via SYs and NTs was not statistically significant, but the loss in precision was statistically significant for AQE via RTs and BTs. (4) SYs and NTs were found to be good candidates for AQE, because they increased relative recall with a loss in precision that was not statistically significant. (5) NTs were found to be good candidates for IQE when precision is desired. (6) RTs were found to be good candidates for IQE because they increased relative recall with a decrease in precision that was less than with AQE.

This study appears to be the first to explore AQE via semantic relationships in the operational setting, and to compare AQE and IQE via semantic relationships in the same system. This study confirms a number results found with research conducted in experimental settings. Furthermore, this study helps to establish a base for further exploration of this topic in the operational environment and for further AQE and IQE comparisons involving semantic relationships. This conclusion of this study identifies future research directions involving semantic relationships that may ultimately assist with information retrieval system design.

MPACT Scores for Jane Greenberg

A = 4
C = 7
A+C = 11
T = 4
G = 1
W = 4
TD = 4
TA = 0
calculated 2017-05-10 22:36:52

Advisors and Advisees Graph

Directed Graph

Students under Jane Greenberg

ADVISEES:
- Abe J. Crystal - University of North Carolina, Chapel Hill (2008)
- Jacob Kramer-Duffield - University of North Carolina, Chapel Hill (2010)
- Erik T. Mitchell - University of North Carolina, Chapel Hill (2010)
- Angela Patricia Murillo - University of North Carolina, Chapel Hill (2016)

COMMITTEESHIPS:
- Mark A. Rosso - University of North Carolina, Chapel Hill (2005)
- Meredith Rachelle Evans - University of North Carolina, Chapel Hill (2006)
- Kathy Wisser - University of North Carolina, Chapel Hill (2009)
- Lewis Church - University of North Carolina, Chapel Hill (2010)
- Alice Sunday Etim - University of North Carolina, Chapel Hill (2010)
- Yaxaio Song - University of North Carolina, Chapel Hill (2010)
- Xi Nui - University of North Carolina, Chapel Hill (2012)