Context-Adaptive Approach for Automated Entity Relationship Modeling

Abstract

Even a smart data modeler may not be an expert in terms of job knowledge. Hence, the design of a database model is limited by the data modeler's resolution and subjectivity. Because the data modeler transforms a domain user's representations into a database model on the basis of arbitrary decisions, the data modeler may distort or lose information. The best way of designing a database is for a domain user to lay out a database, though this approach might impose a heavy modeling burden on the user. Many traditional automated design systems have failed to become widely used. We propose a new model, the association-based conceptual model (ABCM), for an ordinary field worker. The ABCM does not require a user to have expert knowledge to discriminate entities from attributes and relies solely on business descriptions to generate an appropriate ERD. We devise a context-adaptive approach to automate the creation of ERD, which means that ER modeling depends on the context of a business description. Accordingly, this approach performs modeling by analyzing contexts in a business description that the user creates and then utilizing associations among the various contexts. We introduce the scope of the proposed system and present the detailed logic of the system. Finally, we perform a case study to evaluate the devised system's applicability to practical business fields.