Evaluation of Proximity and Classification of Binary Objects Under Uncertainty

Abstract

To solve the problems of proximity assessment, classification and recognition of objects with qualitative features, feature coding using a binary alphabet is often used, and then proximity functions are applied the encoded information Russel and Rao, Sokal and Michener, Jaccard and Needhan etc. However, when coding individual features in real conditions, uncertainties may arise. For example, the color of luminous objects can change continuously. The values of the monitored digital signals in the devices, depending on the modes of their operation, can vary over significant ranges, which can lead to uncertainty in determining their true values and requires the use of alphabets with greater significance to encode them or the use of the mathematical apparatus of fuzzy logic. However, the use of fuzzy logic and multivalued alphabets, in turn, requires modification of the known proximity functions and the development of new computer components for their software implementation.