Building numerical sensitivity analysis systems using a knowledge based approach

Sensitivity analysis provides a good mechanism for experimenting with parameters to achieve desired changes in the behavior of a problem model. The authors have built an interactive system for sensitivity analysis that combines the use of an object-oriented inheritance network for knowledge representation and qualitative reasoning for automated guidance. This system helps a user carry out sensitivity analysis experiments in the domain of financial analysis of acquisitions. The authors describe the architecture of this system, with an emphasis on how the relevant AI technologies used play a critical role towards the usability and power of the system.<
>