Theses and Dissertations from UMD >
UMD Theses and Dissertations >
Please use this identifier to cite or link to this item:
|Title: ||Studying the Relationships of Information Technology Concepts|
|Authors: ||Tsui, Chia-jung|
|Advisors: ||Wang, Ping|
|Department/Program: ||Library & Information Services|
|Sponsors: ||Digital Repository at the University of Maryland|
University of Maryland (College Park, Md.)
co-occurrence analysis, Information technology management, IT classification, KL divergence
|Issue Date: ||2011|
|Abstract: ||Different information technology concepts are related in complex ways. How can the relationships among multiple IT concepts be described and analyzed in a scalable way? It is a challenging research question, not only because of the complex relationships among IT concepts, but also due to lack of reliable methods. Seeking to meet the challenge, this dissertation offers a computational approach for analyzing, visualizing, and understanding the relationships among IT concepts.
The dissertation contains five empirical studies. The first study employs Kullback-Leibler (KL) divergence to compare the semantic similarity of forty-seven IT concepts discussed in a trade magazine over a ten-year period. Results show that the similarity of IT concepts can be mapped in a hierarchy and similar technologies demonstrated similar discourses. The second study employs co-occurrence analysis to explore the relationships among fifty IT concepts in six magazines over ten years. Results show general patterns similar to those found in the first study, but with interesting nuances. Together, findings from the first two studies imply reasonable validity of this computational approach. The third study validates and evaluates the approach, making use of an existing thesaurus as ground truth. Results show that the co-occurrence-based IT classification outperforms the KL divergence-based IT classification in agreeing with the ground truth. The fourth study is a survey of information professionals who help evaluate this computational approach. Results are generally consistent with the findings in the previous study. The fifth study explores the co-occurrence analysis further and has generated IT classifications very much similar to the ground truth.
The computational approach developed in this dissertation is expected to help IT practitioners and researchers make sense of the numerous concepts in the IT field. Overall, the dissertation establishes a good foundation for studying the relationships of IT concepts in a representative, dynamic, and scalable way.|
|Appears in Collections:||UMD Theses and Dissertations|
Information Studies Theses and Dissertations
All items in DRUM are protected by copyright, with all rights reserved.