Detection of Outliers in Binomial Regression Using CERES and Partial Residual Plots

Abstract

Conditional expectations and residuals (CERES) and partial residual (PR) plots have been used in linear regression model for the identification of outliers. But not much work has been done on how they perform in generalized linear models (GLM). Binomial regression model is a very important type of GLM which have wide applications in dealing with Liver cancer and many other types of data. In this paper, CERES and PR plots is used in binomial regression to detect the outliers. Through real data set, the performance of these plots on the detection of possible outliers is observed separately. The CERES plot performs well in order to diagnose this problem. However, the performance of the CERES and PR plot is found very similar to detect outlier by using simulated data but visualization of CERES plot is better as compared to the PR plots.