Identification of Influential Points in a Linear Regression Model

Abstract

The article deals with the detection and identification of influential points in the linear regression model. Three methods of detection of outliers and leverage points are described. These procedures can also be used for one-sample (independent)datasets. This paper briefly describes theoretical aspects of several robust methods as well. Robust statistics is a powerful tool to increase the reliability and accuracy of statistical modelling and data analysis. A simulation model of the simple linear regression is presented.

Keywords

• Leverage points
• outliers
• method of least squares – LSM
• trimmed mean
• confidence interval
• regression model
• data simulation
• robust estimators