Artificial neural networks as a tool of modeling of training loads.

This paper shows that extremely important element of forming speed capabilities is proper (quantitative) structure of exercise loads. This means that training means should be chosen from point of view of energy production in metabolic processes, which depends on the structure of training means from the information area and energy area, therefore on the character of work made, its intensity, duration of exercise, number of repetitions and duration of rest periods. From the training process effectiveness point of view, it is extremely important to find the correct tool for choosing means in given training cycle. The investigation results confirm the experiences of coaches and theorists of sport, that the structure of volume and intensity of exercise loads should be individually chosen with consideration of predispositions of separate athletes. Individualization of training is condition for its optimization.