Login

Publications  •  Project Statistics

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

Dissertation Information for Yun-hua Shih

NAME:
- Yun-hua Shih
- (Alias) Yunhua Shi

DEGREE:
- Ph.D.

DISCIPLINE:
- Library and Information Science

SCHOOL:
- University of California, Berkeley (USA) (1989)

ADVISORS:
- Michael D. Cooper

COMMITTEE MEMBERS:
- Ray R. Larson
- Stephen Harry West

MPACT Status: Fully Complete

Title: Regular expression searching in Chinese information retrieval

Abstract: Chinese character entry has long been recognized among Chinese speakers as a critical problem area in Chinese text editing and information retrieval systems design. This dissertation is a general inquiry into the problems of providing effective techniques to enter and access stored Chinese information. It starts with a description of three fundamental differences in English and Chinese information retrieval. The characteristics of Chinese written and spoken language, and difficulties these characteristics create for character encoding and retrieval are explored. The issue of how to facilitate character entry and retrieval by regular-expression searching is discussed.

A regular expression in text editing and in information retrieval is a search pattern composed of a mixture of symbols and metasymbols. The symbols match exactly the same symbols in the source file; metasymbols on the other hand, have special meanings which are specified by the system designer. Regular expression searching in Chinese character entry and retrieval has the advantages of saving users' search time and mental effort, and reducing mental model mismatch errors. But regular expression searching will also retrieve noisy information because of the metasymbols contained in the regular expression.

The purpose of this dissertation is to develop mathematical models to control noisy information and to do cost-benefit analysis in regular expression searching for Chinese characters or character strings. Four sets of mathematical models are worked out based on the assumption of random naive user searching. These models can be used to fulfill three tasks: (1) analyze the benefit and cost of employing metasymbols in the regular expression and find regular expressions which provide the largest net benefit; (2) for a given amount of acceptable noisy information, find regular expressions which employ the maximum number of metasymbols; and (3) for a given number of metasymbols, determine regular expressions which generate the minimum amount of noisy information. The limitation of the mathematical models and future research in regular expression searching are discussed.

MPACT Scores for Yun-hua Shih

A = 0
C = 0
A+C = 0
T = 0
G = 0
W = 0
TD = 0
TA = 0
calculated 2008-05-29 14:15:54

Advisors and Advisees Graph

generating graph, please reload

Students under Yun-hua Shih

ADVISEES:
- None

COMMITTEESHIPS:
- None