Qualitative Methods for Reasoning Under Uncertainty (Artificial Intelligence)

In this book Simon Parsons describes qualitative methods for reasoning under uncertainty, "uncertainty" being a catch-all term for various types of imperfect information. The advantage of qualitative methods is that they do not require precise numerical information. Instead, they work with abstractions such as interval values and information about how values change. The author does not invent completely new methods for reasoning under uncertainty but provides the means to create qualitative versions of existing methods. To illustrate this, he develops qualitative versions of probability theory, possibility theory, and the Dempster-Shafer theory of evidence.

According to Parsons, these theories are best considered complementary rather than exclusive. Thus the book supports the contention that rather than search for the one best method to handle all imperfect information, one should use whichever method best fits the problem. This approach leads naturally to the use of several different methods in the solution of a single problem and to the complexity of integrating the results problem to which qualitative methods provide a solution.

Simon Parsons is a Associate Professor in the Department of Computer and Information Science at Brooklyn College and the editor of the journal Knowledge Engineering Review.