Dissertation Information for Walt Trybula
- Walt Trybula
- (Alias) Walter Joseph Trybula
- Library and Information Science
- University of Texas, Austin (USA) (1999)
- Ronald E. Wyllys
- Sanda Erdelez
- Mary Lynn Rice-Lively
- Brooke E. Sheldon
- John M. Konopka
MPACT Status: Fully Complete
Title: Text mining and knowledge discernment: An exploratory investigation
Abstract: Data Mining and Knowledge Discovery have been proven as a means of uncovering hidden information in alphanumeric databases. These databases are based on fields in relational or flat file structures. There is a correlation between the fields and the information contained in them. It has been estimated that over 80% of all written material first occurs in electronic form. The emergence of initial efforts in Text Mining appears to be promising a method of uncovering hidden information in textbases (electronic textual repositories). The textbases are a collection of free-form, connected discourse. The implications are that the successful implementation of such a methodology for uncovering hidden information would provide researchers with unsuspected insights into previously published material.
The purpose of this work was to investigate the methodology that defines the steps required for text-mining and knowledge-discernment in textbases and evaluate the existing instruments. This work assessed the state of the field. It provided a broader evaluation of the requirements of text-mining in light of the functionality provided by data-mining to databases. Furthermore, this work investigated the published data on text-mining tools, developed the functionality required for text mining, and evaluated the application of the text-mining tool(s). Testing the functionality of the tool(s) with both a descriptive and a results oriented textbase provided an evaluation of the state of the knowledge discernment technology for textbases. Building on the findings, a recommendation was made for the developmental needs for future tools. The output of this work is a process description of the elements required for text-mining and knowledge-discernment to be applicable to technical information. The results of this work provide guidance for future developmental efforts in text mining.