Uncertain Evidence and Artificial Analysis.
Abstract
Belief function models form a natural basis for the construction of artificial analysts (eg, expert systems) capable of uncertain judgments. Random samples of various types may be used as the uncertain evidence which defines the necessary probability structure for the model. The paper develops some simple examples to illustrate these ideas. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 16, 1987
- Accession Number
- ADP005292
Entities
People
- A. P. Dempster
- Augustine Kong
Organizations
- Harvard University