论文检索

期望频数受线性约束时对数线性模型的最大似然方法AI 解读

Hypotheses relating to log-linear models may be defined in terms of constraints on the frequencies or on the logarithms of the frequencies. A two-step algorithm is developed to allow fitting models containing constraints of both types. This algorithm is shown to yield maximum likelihood estimates of the log-linear parameters. Three examples are presented to illustrate cases in which these models are of interest.

摘要暂缺

二乘二列联表卡方检验的连续性校正比较AI 解读

This article considers the two-sided test of independence in a 2 × 2 contingency table. The chi-squared approximation to the exact significance of Pearson's X 2 test is compared with four approximation methods adjusting for continuity. Computations were carried out for nearly 150,000 fourfold tables. It was found that both the uncorrected chi-squared test and the Yates correction may give misleading outcomes even for moderate sample sizes and expected values, whereas the…

摘要暂缺