• Dr. KESAVULU POOLA
• Prof. M. Bhupathi Naidu

The problem of outliers is very common in nonlinear models and identification of these outliers also complicated. . In this article we propose several outlier detection techniques for nonlinear regression. The main idea is to use the linear approximation of a nonlinear model and consider the gradient as the design matrix. This paper builds the algorithm to compute the Studentized and Predicted residual sum of squares (PRESS) when obtaining nonlinear equations. PRESS is a well-known “leave-one-out” (LOO) cross-validation method. This method is more significant in regression analysis to decide, how well the model predicts for new observations. This paper develops a method to approximate cross-validation statistics for nonlinear regression. The main objective of is to explain the importance of using the Predicted residual sum of squares (PRESS). This paper advocates the concept of cross-validation and recommends using PRESS for cost analysis. And several business statistical packages assume that, PRESS is for linear and log-linear models. But even though we can calculate PRESS directly by definition for a nonlinear equation, we should avoid running nonlinear regression multiple times.

KEY WORDS: Studentized, PRESS, Outliers, Cross-Validation.

"IMPORTANCE OF STUDENTIZED AND PRESS RESIDUALS FOR NONLINEAR MULTIVARIATE REGRESSION MODELS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.5, Issue 7, page no.353-356, July-2018, Available :http://www.jetir.org/papers/JETIR1807400.pdf

"IMPORTANCE OF STUDENTIZED AND PRESS RESIDUALS FOR NONLINEAR MULTIVARIATE REGRESSION MODELS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.5, Issue 7, page no. pp353-356, July-2018, Available at : http://www.jetir.org/papers/JETIR1807400.pdf