A procedure is presented for removing the influence of seasonal variation from an observed time series. The procedure leads to the development of moving average filters which optimally adjust the data based upon their unique characteristics. In this paper optimal adjustments are determined by the requirements of the decision maker.