การสร้างช่วงความเชื่อมั่นของความแปรปรวนแบบประชากรเดียวด้วยวิธีโบเน็ตบูตสแตรปทีสำหรับข้อมูลที่ไม่มีการแจกแจงปกติ

Abstract

บทคัดย่อไม่สมบูรณ์

Confidence interval of one population variance by Bonett bootstrap-t method for non-normal distributions

The purposes of this study were to develop methods to estimate the variance confidence interval by modifying bootstrap-t method with the confidence interval of variance for non-normal distributions by BonettZs in 2006 method and compare performances of the proposed method with other methods. The three methods were the confidence interval of variance, the confidence interval of variance by BonettZs and the confidence interval of variance by BonettZs with bootstrap-t. The results indicated that for all studied distributions the confidence interval of variance by BonettZs with bootstrap-t performed better when the positive skewness distributions except gamma (6,1) that differs from confidence interval of variance ( σ² ) which more effective when population is normally distributed and donZt performed when the positive skewness distributions.