TY - JOUR

T1 - Application of Artificial Neural Networks in the Problem of Searching for Periodic Signals

AU - Topinskiy, V. V.

AU - Baluev, R. V.

PY - 2026/6/1

Y1 - 2026/6/1

N2 - Abstract: A convolutional neural network is constructed to solve the problem of detecting a sinusoidal signal in a non-uniform time series with periodic gaps, also containing Gaussian white noise. It is shown that this model approximates the posterior probability of signal presence, and the neural network model runs significantly faster than classical Bayesian numerical modeling. Some pitfalls of this method are discussed, in particular the problem of specifying initial approximations for the neural network weights, as well as possible development prospects.

AB - Abstract: A convolutional neural network is constructed to solve the problem of detecting a sinusoidal signal in a non-uniform time series with periodic gaps, also containing Gaussian white noise. It is shown that this model approximates the posterior probability of signal presence, and the neural network model runs significantly faster than classical Bayesian numerical modeling. Some pitfalls of this method are discussed, in particular the problem of specifying initial approximations for the neural network weights, as well as possible development prospects.

KW - Bayesian approach

KW - binary cross-entropy

KW - convolutional neural network

KW - periodic signal detection

KW - time series analysis

UR - https://www.mendeley.com/catalogue/dc3ef388-b707-371d-bb3d-335adc6bab77/

U2 - 10.1134/s1063454126700135

DO - 10.1134/s1063454126700135

M3 - Article

VL - 59

SP - 234

EP - 243

JO - Vestnik St. Petersburg University: Mathematics

JF - Vestnik St. Petersburg University: Mathematics

SN - 1063-4541

IS - 2

ER -