The aim of this study is to develop new methods for forecasting time series with data of different frequencies among exogenous factors; forecasting the indicators of the wholesale electricity market in Russia using methods of combining data of different frequencies, including those based on algorithms of convolutional neural networks. The structure of the work is presented in four sections. The first section analyzes methods for forecasting time series with combining data of different frequencies. The second section presents the architecture of a convolutional network that allows the use of data of different frequencies. The third section presents a price model for the wholesale electricity market using data from the Atlas of Russian Energy. The fourth section presents recommendations and main conclusions of the work.
Published on 07/01/23
Submitted on 30/12/22
Licence: CC BY-NC-SA license
Are you one of the authors of this document?