Selective Overview on Single Diagnostics Methods of Outliers in Logistic regression

Abstract

Logistic regression is an important statistical tool for modeling a set of independent variables that effect in a binary response variable. Most practitioners of statistics have used logistic regression in many scientific areas. Unfortunately, they are not aware that the method of estimation logistic regression breaks down in the presence of outlying data point(s) in the original dataset. The objective of this paper is to bring out the attention of respectful researchers in various scientific areas to single logistic regression diagnostics methods, and the effect of the presence of outliers on logistic regression estimates. Real data are considered in this paper and the results show the high performance of diagnostic methods to detect these observations that are affected on the logistic regression estimates. Some graphs are discussed and supported the results of diagnostic method the influence of outlying data point on the fitted logistic model.