ScholarWorks@숙명여자대학교

초록

One of main aspects of time series analysis is to forecast future values of seriesbased on values up to a given time. The prediction interval for future values is usuallyobtained under the normality assumption. When the assumption is seriously violated, atransformation of data may permit the valid use of the normal theory. We investigatethe prediction problem for future values in the original scale when transformationsare applied in ARMA models. In this paper, we introduce the methodology basedon Yeo-Johnson transformation to solve the problem of skewed data whose modellingis relatively dicult in the analysis of time series. Simulation studies show that thecoverage probabilities of proposed intervals are closer to the nominal level than thoseof usual intervals.

키워드