| This dissertation develops a conceptual data model that can efficiently handle huge volumes of data containing uncertainty and are subject to frequent changes. This model can be used to build Decision Support Systems to improve decision-making process. Business intelligence and decision-making in today's business world require extensive use of huge volumes of data. Real world data contain uncertainty and change over time. Business leaders should have access to Decision Support Systems that can efficiently handle voluminous data, uncertainty, and modifications to uncertain data. Database product vendors provide several extensions and features to support these requirements; however, these extensions lack support of standard conceptual models. Standardization generally creates more competition and leads to lower prices and improved standards of living. Results from this study could become a data model standard in the area of applied decisions sciences.;The conceptual data model developed in this dissertation uses a mathematical concept based on set theory, probability axioms, and the Bayesian framework. Conceptual data model, algebra to manipulate data, a framework and an algorithm to modify the data are presented. The data modification algorithm is analyzed for time and space efficiency. Formal mathematical proof is provided to support identified properties of model, algebra, and the modification framework. Decision-making ability of this model was investigated using sample data. Advantages of this model and improvements in inventory management through its application are described. Comparison and contrast between this model and Bayesian belief networks are presented. Finally, scope and topics for further research are described. |
