An Adaptive Test for Detecting Scale Shift at an Unknown Time Point

• 발행기관: 한국자료분석학회
• 저자명: Il-Seong Song
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.12 No.6, 2973~2980쪽, 전체 8쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2010.12.30

영문 초록

We consider the problem of testing the null hypothesis of no change against scale shift at unknown time point in a series of independent observations. An adaptive two-stage procedure is proposed for this testing problem. The first stage is based on order statistics and the second stage is based on rank test statistics. We use transformations of order statistics which are dependent on measures of tail-weight and skewness of distributions. These measures are estimated by quantiles. Adaptive tests are derived on the basis of the asymptotic relative efficiencies of the rank tests. The empirical powers of the adaptive test are compared with some rank tests through Monte Carlo simulation. It is shown by simulation experiments that adaptive rank tests have good power properties over a wide class of underlying distributions.

목차

1. Introduction

2. A Class of Rank Tests and Asymptotic Properties

3. Adaptive Selection Rule

4. Monte Carlo Study on the Procedure

5. Discussion

References

키워드

change point scale shift rank test adaptive test Monte Carlo simulation