Efficiency of Using One-Dimensional Convolutional Layers in a Neural Network on the Example of Classification a Meteorological Station According to Wind Speeds Time Series Data

Abstract

Data analytics using neural networks is a modern trend in scientific research. One of these scientific tasks is the use of artificial intelligence in the time series research and forecasting. The paper considers the results of experiments on the use of one-dimensional convolutional layers in a neural network within the framework of the meteorological (wind speed) time series data classifying task. The improving of the forecast accuracy by adding convolutional layers is shown, which in the considered real problem reaches 9 %. Several variants of architectures for building a model with convolutional layers with an estimate of the accuracy of their prediction are given.