Metrics
10 Record Views
Journal article
Using rotation transformations to maximize the coefficient of determination in simple linear regression models
European Journal of Mathematics and Computer Science, Vol.3(1), pp.66-78
2016
Abstract
In many instances, excessive variation in observed data limits the utility of simple linear regression. We show that a simple rotation of the coordinate axis about the origin reduces
the observed variance in the data, improves estimates of the slope, and increases the coefficient of determination. Furthermore, we provide a modified version of the basic regression model that accommodates a rotation angle and a method for determining the angle that maximizes the coefficient of determination.
Related links
Details
- Title
- Using rotation transformations to maximize the coefficient of determination in simple linear regression models
- Publication Details
- European Journal of Mathematics and Computer Science, Vol.3(1), pp.66-78
- Resource Type
- Journal article
- Publisher
- Progressive Academic Publishing
- Identifiers
- 99381512142806600
- Academic Unit
- Mathematics and Statistics; Hal Marcus College of Science and Engineering
- Language
- English