Influence on Confidence Regions for Regression Coefficients in Generalized Linear Models
研究了广义线性模型中回归系数置信区域的影响问题，给出了诊断方法，用于检查标准误差对个案权重、解释变量或响应变量观测值的敏感性，并通过实例与标准诊断方法比较。
Influence on confidence regions has received less attention than the problem of influence on regression coefficients in generalized linear models. Diagnostics are given for examining influence on confidence intervals or the confidence region for all of the regression coefficients. These diagnostics permit one to check whether the standard errors of the coefficients are sensitive to the exact specification of case weights, or the observed values of the explanatory variables or the response. The diagnostics are illustrated with an example and compared to standard diagnostic methods.