Abstract
Adequacy of the Smirnov approximation to P (D mn ≧c/mn), of the exact distribution of the two sample Kolmogorov-Smirnov criterion,D mn ,m<=n is examined. The main finding is that accuracy depends as much on the sample sizes,m andn, as on the ratio, ρ=m/n. The Smirnov approximation is good whenn is an integer multiple ofm, especiallyn=m, 2m; poor otherwise. This contrasts with the Smirnov approximation to P (D mn >c/mn), where this ordering is reversed, the cases forn=m, 2m being poor. The merit of continuity correction, of order ofn −1 in theD mn scale, is demonstrated for the Smirnov approximation, .1≦ρ, as well as for the Kolmogorov approximation, ρ<.1. As an argument for the optimum choice of ρ, the casen=m+1 is shown to have much to recommend it over the casen=m. Finally the usefulness of theχ 2 distribution with 2 degress of freedom is illustrated as an approximation to the Smirnov distribution.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: Annals of the Institute of Statistical Mathematics
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
