Interpolated Nonparametric Prediction Intervals and Confidence Intervals
Journal of the Royal Statistical Society: Series B (Statistical Methodology) · 1993 · Volume 55 · Issue 3
DOI: 10.1111/j.2517-6161.1993.tb01929.x
Abstract
In several important statistical problems, prediction intervals and confidence intervals can be constructed with coverage levels which are known precisely but cannot be rendered equal to predetermined levels such as 0.95. One solution to this difficulty is to interpolate between such intervals. We show that simple linear interpolation reduces the order of coverage error, but that higher orders of interpolation produce no further improvement. The error is reduced by a factorn−1for prediction intervals andn−1/2for confidence intervals, wherendenotes sample size. In the case of confidence intervals for quantiles, linear interpolation provides particularly accurate intervals which err on the side of conservatism.
Keywords
confidence interval, coverage error, interpolation, median, non-parametric prediction interval, quantile
Enable JavaScript for citation tools and additional paper interactions.