Abstract
This paper presents a method for improving the accuracy of polynomial regression analysis. It is applicable when the relationship being investigated is between one independent and one dependent variable.
The least squares equation which best “fits” the data is determined using the properties of orthogonal polynomials and the statistical technique known as Analysis of Variance. The major differences between this technique and the usual regression approach to curve fitting are also discussed.
Additional information
Notes on contributors
Alan F. Dutka
Mr. Dutka, an ASQC member, is a Management Scientist with the B. F. Goodrich Company in Akron, Ohio, and is a lecturer in the James J. Nance College of Business Administration, Cleveland State University, Cleveland, Ohio. Mr. Ewens is Senior Scientific Programmer with B. F. Goodrich in Akron.