Open AccessNeural network technologies in system synthesis
Author(s) -
О. Филатова,
Yu. V. Bashkatova,
Л. Шакирова,
М. Филатов
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1047/1/012099
Subject(s) - artificial neural network , ranking (information retrieval) , biomedicine , principal (computer security) , artificial intelligence , identification (biology) , computer science , pattern recognition (psychology) , machine learning , biology , bioinformatics , botany , operating system
Work of artificial neural networks does not ensure the identification of order parameters (which are the principal diagnostic characters x i in biomedicine). We suggest to eliminate the 1st type uncertainties (when samplings x i statistically match for different physiological states of a human body) by introducing random setting of initial weight values w io of x i and subsequent multiple repetition ( n ≥1000) of artificial neural network settings. The x i ranking is made according to the weight samplings w i collected after such settings are applied.