least squares
Definition:
This is a method of determining an equation of 'best-fit' to a set of data to enable the prediction of one variable from others. That equation is chosen which minimizes the total of the differences (after squaring them) between each data point and the corresponding mathematical properties, but equations derived from it have no validity beyond the arbitrary choice of the criterion.
Cross-References:
Links:
Figures:
© Westburn Publishers Ltd 2002, The Westburn Dictionary of Marketing edited by Michael J Baker, ISBN 978-0-946433-01-8. www.themarketingdictionary.com. Entry: [J. A. Bound],.