GSLIB Help Page: HISTSMTH
Creates a smooth distribution model constrained to a mean, variance,
quantiles, and smoothmess
datafl: the data file with the raw (perhaps declustered) data.
icolvr and icolwt: the column location for the variable
and the declustering weight (0 if none available).
tmin and tmax: all values strictly less than tmin
and strictly greater than tmax are ignored.
title: a 40-character title for the top of the PostScript plot.
psfl: name for the PostScript output file.
nhist: number of histogram classes for the PostScript output
file. nhist is typically set less than nz (see below) to
obtain a reasonable histogram display superimposed on the smoothed
outfl: output file containing the smoothed distribution (evenly
spaced z values and variable p values).
nz, zmin and zmax: the number N of evenly spaced
z values for the smoothed histogram and the limits for the
evenly spaced z values
ilog: =0 then an arithmetic scaling
is used, =1 then a logarithmic scaling (base 10) is used.
maxpert, report, omin and seed: after
maxpert x nz perturbations the program is stopped.
After report x nz perturbations the program reports on
the current objective function(s). When the normalized objective
function reaches omin the program is stopped.
The random number seed seed should be a large odd integer.
imean, ivari, ismth and iquan: flags for whether
closeness to a target mean, closeness to a target variance,
smoothness, and closeness to specified quantiles will be considered
(1 = yes, 0 = no).
sclmean, sclvari, sclsmth and sclquan: user imposed
weights which scale the weights that the program automatically
nsmooth: half of the smoothing window size.
mean and variance: target mean and variance (if set less
than -999 then they will be calculated from the input data).
ndq: number of quantiles defined from the data (evenly spaced
cumulative probability values).
nuq: number of quantiles defined by the user (nuq lines
must follow with a cdf and a z value). The user-defined quantiles
allows the user more control over ``peaks'' and ``valleys'' in the
smoothed distribution. The user should choose these quantiles to be
consistent with the quantiles defined from the data.