| Quantitative and qualitative information are often complimentary in decision making. While quantitative information processing for decision support has been increasingly automated, the provision of computer integrated qualitative information, mostly represented in texts, seems to lag behind. The information base used for computer decision support can be expanded from the traditional data base, model base, and knowledge base to include a text base. Thus, decision support systems (DSS) can be designed to allow large collections of text to be organized, stored, retrieved, and integrated with quantitative information to provide better decision support. The purpose of this dissertation is to formulate a conceptual DSS design that incorporates a text base, and based on such a design, to develop a workable and testable prototype DSS in a selected application domain, corporate performance analysis. A graphic-oriented, window-operated DSS prototype was developed. This prototype allows a decision process consisting of such tasks as ratio selection, quantitative data retrieval for the ratio, ratio execution, and textual information retrieval to support the ratio analysis to be integrated and supported. Using an intelligent agent (IA) as an interface, the system is also capable of flexibly interacting with users in model construction, automatically determining the location of the financial data needed for ratio execution, and intelligently deriving search keys to retrieve relevant textual information. |
