selfStart model evalueates the Gaussian model and its
gradient. It has an initial attribute that will evalueate
the inital estimates of the parameters mu, sigma,
and h.
SSgauss(x, mu, sigma, h)x. It is the value
of the expression h*exp(-(x-mu)^2/(2*sigma^2), which is a
modified gaussian function where the maximum height is treated
as a separate parameter not dependent on sigma. If arguments
mu, sigma, and h are names of objects, the
gradient matrix with respect to these names is attached as an
attribute named gradient.
mu and h are chosen from the
maximal value of x. The initial value for sigma is
determined from the area under x divided by h*sqrt(2*pi).
nls,
selfStart