You are working with the text-only light edition of "H.Lohninger: Teach/Me Data Analysis, Springer-Verlag, Berlin-New York-Tokyo, 1999. ISBN 3-540-14743-8". Click here for further information. |
See also: variable selection |
C.L. Mallows [1] developed a method to find adequate models by plotting a special statistic against the number of variables+1.
C_{p} = SS_{res}/MS_{res} - N + 2p,
The general procedure to find an adequate model by means of the C_{p} statistic is to calculate C_{p} for all possible combinations of variables and the C_{p} values against p. The model with the lowest C_{p }value approximately equal to p is the most "adequate" model.
[1] | C.L. Mallows
Some comments on Cp Technometrics 15 (1973) 661 |
Last Update: 2004-Jul-03