An Extended Relational Database Model for Uncertain and Imprecise Information
Suk Kyoon Lee · Very Large Data Bases · 1992
We propose an extended relational database model which can model both uncertainty and imprecision in data. This model is based on Dempster-Shafer theory which has become popular in AI as an uncertainty reasoning tool. The definitions of Be1 and Pls functions in Dempster-Shafer theory are extended to compute the beliefs of various comparisons (e.g., equality, less than, etc.) between two basic probability assignments. Based on these new definitions of Be1 and Pls functions and the Boolean combinations of Be1 and Pls values for two events, five relational operators such as Select, Cartesian Product, Join, Projection Intersect, and Union are defined.