hohle and hohle(2009) claim that the previous assumption of a gaussian distribution of ดาต้า with few outliers is not valid for numerous dem; thus, accuracy measure have to take into account that outliers may exist and that the distribution of errors might not be normal. they introduce the measure of the absolute mean error (ame) as the dem evelution is interested. one approach to deal with outliers is to remove them by applying a threshold. hohle and potockova (2006) select a threshold for eliminating outliers of three times the rmse. the rmse value is calculated from an initial assessment of the deta data. this approach classifies a deta value as an outlier if dets (maune...)despite these methods not all of the outlier can be identified and the accuracy measure might therefore be flawed.