One-step Prediction of Air Pollution Control Parameters using Neural-Like Structure Based on Geometric Data Transformations

The article has aimed at the investigation of a one-step prediction by neural-like structures using the Model of Geometric Data Transformations. The work describes the timestamp method for a one-step prediction. It has indicated that the smaller the time window for the selected parameters for environmental monitoring, the more accurate the prediction error. Moreover, in the article is proved that non-iterative linear neural-like structures of Geometric Data Transformations do better one-step prediction than Radial Bases Function neural network and linear neural-like structure combined with Radial Bases Function.