©
Nelder-Mead regression
  Applies the Nelder & Mead minimization method to a regression problem.  
2024.May.17 11:45:10
n, nvar No. of experimental points and independent variables.
npar No. of adjustable parameters.
. criterion 1 2 (min. sq.) 3 4 ∞ (minmax) Criterion [power of |ycalc-y|]. •
X x-values [column-wise, ie., x(:, j), j = 1, nvar, ie., 'nvar' rows]
•
Y y-values
•
P p-values, adjustable parameters
•
tol, maxite, monit Tolerance, max. no. of iterations, monitoring. •
Show values Show the values of the graph coordinates.
  Applies the Nelder and Mead [1965] minimization method to a regression problem (here a DOE-type polynomial [Myers et al., 2002]). A total discrepancy, the (unweighted) sum of the deviations between the calculated response, ycalc, and the experimental value, y, is minimized according to the given criterion (sum of the absolute deviations, each raised to the user given power, then with the inverse power applied).
  A graph will be drawn for the experimental and calculated points.
References: Plate NM08504

• Nelder, J. A., and R. Mead, 1965, "A simplex method for function minimization", Computer Journal, 7, pp 308–313 (National Vegetable Research Station, Wellesbourne, Warwick, UK).

• Myers, Raymond R. and Douglas C. Montgomery, 2002, "Response surface methodology: process and product optimization using designed experiments", 2.nd ed., John Wiley & Sons, Inc., New York, NY (USA), xv+798 pp, ISBN 0-471-41255-4 (IST).

• Weisstein, Eric W., "Nelder-Mead method". From MathWorld–a Wolfram Web Resource [on "created" date].

• 1733-05-04: Borda, Jean Charles de.

 
 
Valid HTML 4.01! IST http://web.ist.utl.pt/~mcasquilho/compute/com/Fx-nelmea.php
Created: 2008-05-04 — Last modified: 2017-04-26