The effect of methods of eliminating spikes in the time series of freight flows on their statistical characteristics

Abstract

Types of the data aggregate spikes of the time series were analyzed taking as an example the incoming freight flows at the industrial enterprise. Effect of the method of correction of abnormal series levels on its quantitative characteristics and the trend component was established. The ranges of deviation from the mean value of actual data were determined. Decrease in dispersion for one-step correction up to 20 % and up to 99 % for iterative correction was determined. It was established that the degree of correlation of the time series levels weakly reacts to the method of correction of abnormal values. Methods of elimination of the trend and cyclic components by iterative correction of abnormal observations by means of different estimators were considered. As a result of partial robust processing of abnormal values, an updated time series was obtained/ This time series can be used further for modeling and predicting indicators studied for different systems. It was established that the remaining parts of the deterministic actual series accumulate in themselves about 75 % of dispersion of the actual series for one-step correction and 54 % for iterative correction. On average, 6 % of the actual series dispersion is the share of the trend for all methods of correction (except VOS and MO).