Conference proceeding
Semi-automatic Conceptual Data Modeling Using Entity and Relationship Instance Repositories
CONCEPTUAL MODELING - ER 2011, v 6998
01 Jan 2011
Abstract
Data modelers frequently lack experience and have incomplete knowledge about the application being designed. To address this issue, we propose new types of reusable artifacts called Entity Instance Repository (EIR) and Relationship Instance Repository (RIR), which contain ER modeling patterns from prior designs and serve as knowledge-based repositories for conceptual modeling. We explore the development of automated data modeling tools with EIR and RIR. We also select six data modeling rules used for identification of entities in one of the tools. Two tools were developed in this study: Heuristic-Based Technique (HBT) and Entity Instance Pattern WordNet (EIPW). The goals of this study are (1) to find effective approaches that can improve the novice modelers' performance in developing conceptual models by integrating pattern-based technique and various modeling techniques, (2) to evaluate whether those selected six modeling rules are effective, and (3) to validate whether the proposed tools are effective in creating quality data models. In order to evaluate the effectiveness of the tools, empirical testing was conducted on tasks of different sizes. The empirical results indicate that novice designers' overall performance increased 30.9 similar to 46.0% when using EIPW, and increased 33.5 similar to 34.9% when using HBT, compared with the cases with no tools.
Metrics
Details
- Title
- Semi-automatic Conceptual Data Modeling Using Entity and Relationship Instance Repositories
- Creators
- Ornsiri Thonggoom - Drexel Univ, iSch, Philadelphia, PA 19104 USAIl-Yeol Song - Drexel Univ, iSch, Philadelphia, PA 19104 USAYuan An - Drexel Univ, iSch, Philadelphia, PA 19104 USA
- Contributors
- M A Jeusfeld (Editor)L Delcambre (Editor)T W Ling (Editor)
- Publication Details
- CONCEPTUAL MODELING - ER 2011, v 6998
- Series
- Lecture Notes in Computer Science
- Publisher
- Springer Nature
- Number of pages
- 14
- Resource Type
- Conference proceeding
- Language
- English
- Academic Unit
- Information Science
- Web of Science ID
- WOS:000303359500017
- Scopus ID
- 2-s2.0-80455123852
- Other Identifier
- 991019173723904721
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Web of Science research areas
- Computer Science, Artificial Intelligence
- Computer Science, Software Engineering
- Computer Science, Theory & Methods