To obtain access to full text of journal and articles you must register!
- Article name
- Testing the hypothesis of independence of small samples: reproduction of the effects of neurodynamics through random thinning of the source data
- Authors
- Sulavko A. E., , sulavich@mail.ru, Omsk State Technical University, Omsk, Russia
Choban A. G., , adil_choban@mail.ru, Omsk State Technical University, Omsk, Russia
Ivanov А. I., , bio.ivan.penza@mail.ru; ivan@pniei.penza.ru, Penza Research Electrotechnical Institute, Penza, Russia
Zolotareva T. A., , sulavich@mail.ru, Lipetsk Pedagogical University named after P. P. Semenova-Tyan-Shanskogo, Lipetsk, Russia
- Keywords
- artificial neurons / criteria of independence hypothesis verification / small samples / modulation of neuron input data
- Year
- 2020 Issue 4 Pages 42 - 47
- Code EDN
- Code DOI
- Abstract
- The article deals with the analysis of small samples on the basis of several statistical criteria for checking the hypothesis of independence, as direct calculation of correlation coefficients by Pearson formula gives unacceptably high error. It is proposed to replace each of the classical statistical criteria for testing the hypothesis of independence with an equivalent artificial neuron. Training of the neuron is carried out based on the condition of obtaining equal probability of first and second kind errors. In addition, to increase the level of reliability of decisions made by the neural network, it is proposed to use modulation by thinning and rearranging the data of the initial small sample. This makes it possible to observe spectral lines of the Humming distance of the artificial neuron network sequence output codes.
- Text
- BUY for read the full text of article
- Buy
- 500.00 rub
