On Test Statistics for Detecting Outliers in Gamma Distribution

• 발행기관: 한국자료분석학회
• 저자명: Shin Young Park Dong Wook Kim Dong Hee Kim
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.14 No.5, 2295~2304쪽, 전체 10쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2012.10.30

영문 초록

Recently, many authors have discussed test statistics for testing an outlier in exponential distribution, which are Dixon’s statistic, Kumar and Lalitha’s statistic. This distribution is just a special case of the Gamma distribution. In this paper, we propose the test statistic for detecting outliers in Gamma distribution and induce the exact distribution. Throughout the simulation study, we compare the critical values of our proposed test statistics, containing other comparative test statistics. From the simulation results, we show that as sample size n increases, the critical values of our proposed statistics increase.

목차

1. Introduction

2. The Proposed Statistic and Its Distribution

3. Critical Values of the Test Statistics

4. Conclusion and Further Works

References

키워드

Detecting outliers Gamma distribution Dixon’s statistic Kumar and Lalitha’s statistic