Login

Publications  •  Project Statistics

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

Dissertation Information for Seonghee Kim

NAME:
- Seonghee Kim

DEGREE:
- Ph.D.

DISCIPLINE:
- Library and Information Science

SCHOOL:
- University of Wisconsin, Madison (USA) (1994)

ADVISORS:
- Edwin-Michael Cortez

COMMITTEE MEMBERS:
- Douglas L. Zweizig
- Richard D. Walker
- Sang-Chan Park

MPACT Status: Fully Complete

Title: Intelligent information retrieval using an inductive learning algorithm and a back-propagation neural network

Abstract: This study demonstrated that the neural network inductive learning model (NNILM) can retrieve relevant documents from incomplete queries. In addition, this study showed that the neural network inductive learning model outperformed a vector space model even though incomplete queries were used with the NNILM. It presented a design for the application of inductive learning to information retrieval systems, in which an inductive algorithm was merged with a neural network to create a new information retrieval model. The performance results of this neural network inductive learning model were measured by comparing searches in response to complete queries with searches in response to incomplete queries in terms of three meausures: (1) the total number of relevant documents retrieved, (2) precision ratios, and (3) recall ratios. Furthermore, in order to demonstrate the predicted superiority of the neural network inductive learning model, its effectiveness in response to incomplete queries was compared to the effectiveness of the vector space model in response to complete queries. ADI (American Documentation Institute) documents and queries were selected to test the proposed model. The collections consist of 82 documents and 35 queries on the subject of Library and Information Science.

Major findings were as follows:

The results of searches in response to complete queries in a neural network inductive learning model weren't found to be significantly different from those made in response to incomplete queries when measured in terms of the performance criteria. In addition, this study demonstrated that a neural network inductive learning model outperformed a vector space model. These results confirm the usefulness of intelligent information retrieval and encourage the creation of a new information retrieval system having the flexibility to respond more accurately to individual preferences, changing information needs, and different retrieval situations even though the user

MPACT Scores for Seonghee Kim

A = 0
C = 0
A+C = 0
T = 0
G = 0
W = 0
TD = 0
TA = 0
calculated 2008-01-31 06:27:26

Advisors and Advisees Graph

generating graph, please reload

Students under Seonghee Kim

ADVISEES:
- None

COMMITTEESHIPS:
- None