Abstract
Rank regression, which is quite simple to use some form of monotonic
relationship between X and Y. Since the rank regression is a nonparametric
approach so there are essentially no confidence interval, hypothesis tests,
prediction intervals, and interpretation of regression coefficients. In this article,
we proposed a bootstrap statistical significance measure of the rank regression by
formulating a bootstrap interval for the rank regression parameters. If the rank
regression parameters from the original data are not within the bootstrap interval,
the rank regression parameters are considered significance. Numerical examples
show that the merit of using this proposed bootstrap interval.