Separation of the refined estimator of the measure for symmetry in square contingency tables

被引:1
|
作者
Tahata, Kouji [1 ,2 ]
Auchi, Ryohei [1 ]
Ando, Shuji [1 ]
Tomizawa, Sadao [1 ]
机构
[1] Tokyo Univ Sci, Fac Sci & Technol, Dept Informat Sci, Noda, Chiba, Japan
[2] Tokyo Univ Sci, Fac Sci & Technol, Dept Informat Sci, Noda, Chiba 2788510, Japan
基金
日本学术振兴会;
关键词
Conditional symmetry; Decomposition; Global symmetry; Unbiased estimator; DEPARTURE; ASYMMETRY;
D O I
10.1080/03610918.2023.2231184
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Measures that represent the degree of departure from symmetry, global symmetry, and conditional symmetry have been proposed to analyze square contingency tables. Additionally, a refined estimator of the symmetry measure has been suggested. This study provides refined estimators of the measures for global symmetry and conditional symmetry, and shows the relationship between the proposed estimators and the refined estimator. The proposed estimators can be obtained by the second-order term in the Taylor series expansion. Simulations studies demonstrate that as the sample size increases, the proposed estimators approach the true values faster than the former estimators.
引用
收藏
页码:6274 / 6290
页数:17
相关论文
共 50 条