Answers are grounded in this paper and cite their sources.
This paper presents a time-saving method for computing an iterated moving average with triangular weights to better represent the trend-cycle elements in a time-series for measuring seasonal variation. A printing-adding machine is used to obtain progressive summation of the Y-data values of a time series. A stencil having three suitably spaced slots is used to disclose progressively the three terms which are summed to obtain the graduated series required. Multiplied by the reciprocal of the weights the graduated series is reduced to the moving average desired. The moving average is extended to both ends of the series by means of a truncated weight function. Computation of the seasonal indexes is otherwise conventional. Tested against a synthetic series with known seasonals and the Census Bureau test series for seasonal computations, the iterated moving-average method is found to be superior to others now in use.
No takes yet. Share an insight, caveat, or question.
Y. S. Leong (1962) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: