Combined models for forecasting the air pollution level in infocommunication systems for the environment state monitoring

Abstract

The combined models of selective and hybrid types with indexation of time series for forecasting the level of air pollution in infocommunication systems for monitoring the state of the environment are developed. Indexing in these models takes place on the basis of the nearest neighbor's method with selected metric distances. Described models allow to achieve a higher precision of short-term and medium-term forecasting compared to the models included in the base set of these combined models. Models and appropriate methods can be used in the development of info communication systems for monitoring the state of the environment and hardware-software complexes of general environmental monitoring.