Multiple prediction intervals for time series: Comparison of simultaneous and marginal intervals

Simultaneous prediction intervals for forecasts from time series models that contain L (L ≤ 1) unknown future observations with a specified probability are derived. Our simultaneous intervals are based on two types of probability inequalities, i.e. the Bonferroni- and product-types. These differ from the marginal intervals in that they take into account the correlation structure between the forecast errors. For the forecasting methods commonly used with seasonal time series data, we show how to construct forecast error correlations and evaluate, using an example, the simultaneous and marginal prediction intervals. For all the methods, the simultaneous intervals are accurate with the accuracy increasing with the use of higher-order probability inequalities, whereas the marginal intervals are far too short in every case. Also, when L is greater than the seasonal period, the simultaneous intervals based on improved probability inequalities will be most accurate.

[1]  R. J. Bhansali,et al.  Asymptotic Mean-square Error of Predicting more than One-step ahead Using the Regression Method , 1974 .

[2]  Everette S. Gardner,et al.  A Simple Method of Computing Prediction Intervals for Time Series Forecasts , 1988 .

[3]  D. Hoover,et al.  Subset complement addition upper bounds - an improved inclusion-exclusion method , 1990 .

[4]  Samuel Karlin,et al.  Total Positivity Properties of Absolute Value Multinormal Variables with Applications to Confidence Interval Estimates and Related Probabilistic Inequalities , 1981 .

[5]  W. C. Guenther,et al.  Analysis of variance , 1968, The Mathematical Gazette.

[6]  Spyros Makridakis,et al.  Confidence intervals: An empirical investigation of the series in the M-competition , 1987 .

[7]  R. Shumway,et al.  AN APPROACH TO TIME SERIES SMOOTHING AND FORECASTING USING THE EM ALGORITHM , 1982 .

[8]  D. Hoover.  Comparisons of improved bonferroni and sidak/slepian bounds with applications to normal markov processes , 1990 .

[9]  Eugene Seneta.  DEGREE, ITERATION AND PERMUTATION IN IMPROVING BONFERRONI-TYPE BOUNDS , 1988 .

[10]  K. Worsley.  An improved Bonferroni inequality and applications , 1982 .

[11]  M. Schervish.  Multivariate normal probabilities with error bound , 1984 .

[12]  J. Glaz,et al.  Probability Inequalities for Multivariate Distributions with Dependence Structures , 1984 .

[13]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[14]  Fred M. Hoppe,et al.  Iterating bonferroni bounds , 1985 .

[15]  Arnold L. Sweet,et al.  Computing the variance of the forecast error for the holt‐winters seasonal models , 1985 .

[16]  David J. Hunter.  An upper bound for the probability of a union , 1976, Journal of Applied Probability.

[17]  Nalini Ravishanker,et al.  Approximate simultaneous prediction intervals for multiple forecasts , 1987 .

[18]  Douglas C. Montgomery,et al.  Forecasting and time series analysis , 1976 .

[19]  Johannes Ledolter,et al.  Statistical methods for forecasting , 1983 .

[20]  E. Seneta,et al.  Bonferroni-type inequalities , 1987, Advances in Applied Probability.

[21]  Chris Chatfield,et al.  Holt‐Winters Forecasting: Some Practical Issues , 1988 .

[22]  R. Shumway.  Applied Statistical Time Series Analysis , 1988 .