loess2

Surface fit using local regression. Note: If x and y are not on the same scale, they should be standardized first.

Input Arguments:

• X, Y, and Z are vectors containing the observed data.
• NEWX and NEWY is the domain over which to evaluate the smooth surface.
• ALPHA is the smoothing parameter, ranges between 0 and 1.
• LAMBDA is the degree of the local fit. Use 1 for linear and 2 for quadratic.
• If FLAG is present, then a robust fit is used.

Output Arguments:

• g is the fitted points at the newx, newy locations..

Synopsis

g = loess2(x,y,z,newx,newy,alpha,lambda,flag)

References:

Cleveland, W. S. 1993. Visualizing Data, New York: Hobart Press.

Copyright Notice

Copyright (c) 1998 by Datatool, See the homepage here.