A principal distinction of neural network based methods of approximation from conventional methods is a hidden algorithmic function inside, which is defined by a neural network topology and is enough sophisticate constructed to make appropriate fitting process. There is new construction named a...

High uncertainty of AI formalization leads to misunderstanding and wrong activity. Finite number of words, being implemented somewhere else, creates unacceptable scientific environment, which leads to wrong conclusions. The problem of correct terminological extension from common area to the very...

Traditional forecast in technical analysis, based on time series extrapolation, should be substituted by more generic for human beings forecast by repeated situations. This means circadian and circannual rhythms and any typical behavior, which is analyzed and stored in “whole life” memory....

Известно, что классический метод обучения искусственной нейронной сети, построенной по принципу перцептрона Розенблатта, является модификацией метода решения...