International Science Community AssociationResearch Journal of Mathematical and Statestical Sciences2320 - 60477120190112Statistical diagnostics of models for binomial response16ENChooAli 1School of Science, Department of Probability and mathematical Statistics Nanjing University of Science and Technology, Nanjing, Jiangsu Province 210094, P.R., ChinaFengYu 2School of Science, Department of Probability and mathematical Statistics Nanjing University of Science and Technology, Nanjing, Jiangsu Province 210094, P.R., ChinaAliZamir 3School of Science, Department of Probability and mathematical Statistics Nanjing University of Science and Technology, Nanjing, Jiangsu Province 210094, P.R., ChinaAli Munsir 4School of Science, Department of Probability and mathematical Statistics Nanjing University of Science and Technology, Nanjing, Jiangsu Province 210094, P.R., China>
Research Article2018102220190112In regression analysis, outliers are frequently encountered. Diagnostics of outliers is an essential tool of the model building process. Most of the time analysts depend on ordinary least square (OLS) method without identifying outliers. It is evident that OLS utterly fails in the identification of outliers. In this paper, we use diagnostics techniques to detect residuals and influential points in statistical models for binomial the response. Gauss-Newton and likelihood distance methods were considered to identify the outliers in parameter estimation in non-linear regression analysis. The results illustrated single and multiple outliers in dataset.Copyright@ International Science Community Association