(НТУ "ХПІ", 2016) Kovalenko, I. I.; Chernova, L. S.
This paper proposes a modified moving average method. The basis of the method is to find an effective average estimator on the basis of moving that consists of some subset of the elements of the average series. To improve accuracy of the obtained forecast values the averages test for efficiency at each step of moving is done by the resampling method. This method is actively used in technical and economic analysis, as it has a profound statistical justification. The obtained forecast error values are acknowledged as possessing "satisfactory accuracy" and "good accuracy". Accordingly, the modified method has advantages over other modifications of the moving average method. In future studies of the proposed method in different time series, for example, with so-called "suspicious", "outlier" values the new results can be obtained.
