CORRECTING PREDICTION PROCEDURES WHEN CONTROLLING POLLUTANT EMISSIONS OF ENERGY ENTERPRISES.

The problem of obtaining information redundancy contained in random, generally nonstationary by the mathematical expectation dynamic processes under control was considered in the paper. A mathematical model of predicting critical emissions in random atmosphere pollution processes with the correction to asymmetry of the processes was proposed. The procedure of testing a multicomponent pollution process by parameters during a multivariable control considering its extrapolated values was described. An algorithm of a parametric significance test concerning stationarity violations with regard to risks of the first type control was developed. The test implements a procedure of verifying statistical hypotheses. This approach allows using the extrapolation model, adjusted towards the mathematical expectation, reducing risks of the second-type control. The research results, described in this paper can be used by specialists in the field of ecological control and monitoring, as well as in information systems and technologies. [ABSTRACT FROM AUTHOR]