Deletion Diagnostics in Logistic Regression

Today, there are not many good measures for detecting influential observations in case of fitting a logistic regression model. So, the purpose this article is to extrapolate from pre-existing deletion diagnostics defined points multiple linear regression, i.e. DFFITS, DFBETAS and Cook's Distance scenario binary model then view multinomial as special same. The threshold determining whether an observation or judged using asymptotic distribution setting, both single group deleted case. results examined under various simulation scenarios well over modified Kyphosis data-set.