This paper examines the impact on decision-making of the completeness of data about an object, which can vary from incomplete to redundant. The concept of conditionally redundant data is introduced. As an example, an algorithm implemented in an expert system for the differential diagnosis of rare diseases is presented. The decision-making algorithm enables the ranking of hypotheses by accounting for the presence of conditionally redundant patient information and the degree of similarity with reference descriptions in the knowledge base in the case of incomplete descriptions.
Blagosklonov et al. (Mon,) studied this question.