Knowledge Discovery For Predicting Entity Relationship Diagram Maintainability

It is generally accepted that the quality of the information system (IS) is highly dependent on decisions made early in the IS development. Conceptual data model is a key artifact built at the early phases of the IS life cycle, therefore its quality influence on the quality of IS which is finally implemented. We focus this paper on the maintainability of conceptual data models, because it is one of the most crucial quality characteristic. As maintainability is an external quality attribute that can only be measured once an IS is finished or nearly finished, our idea is to present a set of measures for measuring the structural complexity of Entity Relationships diagram (ERD), whose values can be obtained at the early phases of the IS life cycle, and based on these metrics values we will be able to predict ERD maintainability. For building the prediction model, we have used an extension of the original Knowledge Discovery in Databases (KDD): the Fuzzy Prototypical Knowledge Discovery (FPKD) that consists of the search for fuzzy prototypes that characterise the maintainability of an ERD. These prototypes lay the foundation of the prediction model that will lead us to predict ERD maintainability.