A framework for computer-supported interpretation systems

In recent years organizational computing has received a great deal of attention from both computer scientists and organizational scientists because of the increasing strategic importance of information technology in an organization's success. Computer systems are expected to play more important roles in supporting the ongoing activities of organizations and therefore, expected to acquire more of the characteristics of organizations. A few examples are (group) decision support systems, executive information systems, computer-supported cooperative work, and negotiation support systems. This dissertation proposes a database framework for computer-supported interpretation systems (CSIS), based on the model of organizations as loosely-coupled interpretation systems, to support organization's information interpretation process. We introduce an extensional approach to database management emphasizing flexibility in information acquisition and interpretation. In an extensional database, objects and classes are loosely coupled so that objects can be defined without being a member of a predefined class, the objects in a class need not be homogeneous in their attributes, and objects can be classified inductively based on expert judgement and experience as well as deductively based on structures and rules. We define the Extensional Object Model (ExOM) as a formalism for extensional databases. The ExOM incorporates imprecise data description, exemplar-based concept representation, and machine learning with conventional object-oriented models. Both the deductive and inductive approaches are integrated for interpretation. Thus we explore the possibilities of knowledge discovery or inductive learning, as well as the deductive capabilities, in a database framework. In our view, CSIS will become the base of the next generation organizational information systems with stability and adaptability, providing the capabilities envisioned in organization theories.