Premium
Stopping Distance from Observations by Stepwise Regression
Author(s) -
Niessner Herbert
Publication year - 2014
Publication title -
pamm
Language(s) - English
Resource type - Journals
ISSN - 1617-7061
DOI - 10.1002/pamm.201410364
Subject(s) - regression , stepwise regression , simple (philosophy) , regression analysis , rank (graph theory) , mathematics , symmetry (geometry) , linear regression , statistics , combinatorics , philosophy , geometry , epistemology
In a regression equation with many terms stepwise regression allows to pick out significant terms and to exclude insignificant ones. This can be used to solve rank deficient equation systems. Yet significance is not condition. There is a rather simple method for stepwise regression starting from extended normal equations proposed by Efroymson 1960 [1] and improved by Breaux 1968 [2]. Numerically more stable methods apply orthogonal transformations (Eldén 1972 [3], Gragg‐LeVeque‐Trangenstein 1979 [4]). If symmetry is properly made use of [2, 5] the method of Efroymson is the most efficient one with respect to computing time. But reasonable results are obtained only if the terms in the regression equation are physically justified. This illustrates the search for a formula of the stopping distance derived from observations, an exercise given in [9], a book widely used on some universities. (© 2014 Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim)