Abstract
Applying neural networks, genetic and immune algorithms enable solving tasks of computed-aid diagnosis. In this paper we describe a system for medical data processing that mainly uses neural networks and genetic algorithms. The article provides the results achieved by processing different data including databases of children with perinatal neural system damage. It reveals the profits of using bionic algorithms not only from a medical point but also includes comparison of neural networks architectures and training method variations and their influence on percentage of correctly classified inputs.