On Coverage Probability of the Prediction Interval for Normal Variable

This paper presents a coverage probability of a one-step-ahead prediction interval for a normal variable. This coverage probability is proved to be functionally independent of (µ,ơ²). Because of this functional independence, Monte Carlo simulation will include some variance reduction by setting (µ,ơ²) = (0,1). This result will then valid for all possible values of (µ,ơ²). This leads to a great reduction in computation effort.

Keywords: Coverage probability, prediction interval

Corresponding author: E-mail: snw@kmitnb.ac.th