Using data quality measures in decision-making algorithms

Four decision methods are compared to determine appropriate ways of using data quality measures. Separate studies are directed toward defining and measuring tactical data quality, and calibrating the measures to decision problems. The decision methods compared include Dempster's rule, the linear and logarithmic opinion pools, a fuzzy-logic algorithm and the Mycin certainty factor calculus. It is concluded that none of the four decision algorithms are fully satisfactory. Of the four algorithms, the linear opinion pool is the most likely to succeed in practice because it is the simplest.<<ETX>>