Abstract: An effective mathematical method of new knowledge obtaining on the structure of complex objects with required properties is developed. The method comprehensively takes into account information on the properties and relations of primary objects, composing the complex objects. It is based on measurement of distances between the predicate groups with some interpretation of them. The optimal measure for measurement of these distances with the maximal discernibleness of different groups of predicates is constructed. The method is tested on solution of the problem of obtaining of new compound with electro-optical properties.

Keywords: New knowledge, Predicates, Complex objects, Primary objects, Maximal discernibleness.

ACM Classification Keywords: I.2.4 Artificial Intelligence: knowledge representation formalisms and methods.

Link:

NEW KNOWLEDGE OBTAINING IN STRUCTURAL-PREDICATE MODELS OF KNOWLEDGE

Valeriy Koval, Yuriy Kuk

http://www.foibg.com/ijita/vol12/ijita12-1-p10.pdf