Zaitsev S. Methods of artificial immune system construction in intelligent diagnosis systems

The aim of the thesis is artificial immune system construction methods to enhance their accuracy and training speed, to decrease their complexity and increase generalization level. Research object is the artificial immune system construction process. Research subject is the inductive predicate learning methods of the artificial immune systems. In the thesis the following results have been achieved: a new artificial immune system construction method based on the masked detectors, a new artificial immune system construction method based on the probability distribution of the diagnostic signs; an improved artificial immune system construction method based on the hierarchy structure for recognition; an improved artificial immune network synthesis method for cluster analysis; the methods implemented in software significantly improved the synthesis speed and the accuracy of the recognition models based on immunocomputing, that proves the reasonability of the practical methods usage. The scientific and practical results of work achieved have been introduced in enterprises and used in the educational process.