browse.title.page

Зараз показуємо 1 - 1 з 1

Use of neural networks to maximize the effectiveness of shot putters training
(2021) Kadatskyi, M.
The main factors influencing the results of shot put are considered. The necessity of using modern methods for solving forecasting problems is substantiated. The method of artificial neural networks with different architecture is proposed to solve the following problems: finding the percentage of correction of the shot put technique, finding the subtype of the technique, the activation function of the sigmoid and the algorithm of reverse propagation of errors for learning networks. A software package has been developed that allows to find the approximate result of pushing the shot using the technique of "glide" and "from the ground". Examples of calculations in the environment Deductor Studio Lite are given.