{smcl}
{* 10may1999}{...}
{hline}
help for {hi:varwiker}{right:(SJ3-2: st0036)}
{hline}

{title:Variable bandwidth Gaussian kernel density estimation}

{p 12 19 2}
{cmd:varwiker} {it:varname} [{cmd:if} {it:exp}] [{cmd:in} {it:range}]{cmd:,}
                 {cmdab:b:width(}{it:#}{cmd:)} [{cmdab:g:en(}{it:denvar}{cmd:)} 
                 {cmdab:nog:raph} {it:graph_options}]


{title:Description}

{p 4 4 2}{cmd:varwiker} estimates the density of {it:varname} using the
variable bandwidth Gaussian kernel described in Fox (1990) modified from
Silverman (1986) and draws the result.


{title:Options}

{p 4 8 2}{cmd:bwidth(}{it:#}{cmd:)} permit to specify the width of the window
around each data point.  {cmd:bwidth()} is not optional. If the user does not
provide it, the program halts and displays an error message on screen.

{p 4 8 2}{cmd:gen(}{it:denvar}{cmd:)} permits to generate the variable
{it:denvar} with the density values 

{p 4 8 2}{cmd:nograph} suppresses the graph drawing.

{p 4 8 2}{it:graph_options} are any of the options allowed with {cmd:graph,}
{cmd:twoway}.


{title:Remarks}

{p 4 4 2}{cmd:varwiker} estimates densities using a Gaussian kernel with fixed
window, then uses these estimates to determine local weights inversely
proportional to the preliminary density estimate.  These local weights are
used to adjust the window width so that it is narrower at high densities
(retaining detail) and wider where density is low (eliminating noise).

{p 4 4 2}As with the fixed window kernels the "smoothness" of the result can
be regulated by changing the width of the interval (i.e., varying 'h'). Wide
intervals produce smooth results; narrow intervals will result in noisier
density values.

{p 4 4 2}WARNING: Because this implementation requires the calculation of
local weights for each individual observation based on a preliminary density
estimation, the time required is proportional to _N.  Please be patient. 


{title:Example}

{p 4 8 2}{cmd:. varwiker infmorat,  b(20) gen(vdensity)}

{p 4 4 2}After some time (depending on the number of observations) the results
are plotted and placed in the new variable {cmd:adensity}.

{p 4 4 2}The density values can be used to repeat the graph by typing:

{p 4 8 2}{cmd:. graph vdensity infmorat, c(s) s(.)}
 
{p 4 4 2}The graphic with the density estimates for each original value
("smoothed" histogram) will appear on screen. You can combine this with the
oneway and box options.


{title:References}

{p 4 8 2}Chambers, J. M., W. S. Cleveland, B. Kleiner, and P. A. Tukey. 1983.
	{it:Graphical Methods for Data Analysis}. Belmont, CA: Wadsworth &
	Brooks/Cole.

{p 4 8 2}Silverman, B. W. 1986. {it:Density Estimation for Statistics and Data 
	Analysis}. London: Chapman & Hall.

{p 4 8 2}Fox, J. 1990. Describing univariate distributions. In 
       {it:Modern Methods of Data Analysis}, eds. J. Fox and J. S. Long, 
       58--125. Newbury Park, CA: Sage Publications.


{title:Author}

        Isaias H. Salgado-Ugarte,
        Facultad de Estudios Superiores Zaragoza, U.N.A.M.
        Biologia (Fax 52-5 773-1183)
        {browse "mailto:isalgado@servidor.unam.mx":isalgado@servidor.unam.mx}


{title:Also see}

    {p 5 19 2}STB: snp6 (STB-16){p_end}
{p 5 19 2}Online: help for {help boxdetra}, {help boxdetr2}, {help cosdetra}, 
{help kernsim}, {help kernepa}, {help kerngaus}, {help adgaker2}, 
{help boxdent}, {help dentrace}, {help kerneldm}, {help warpdenm}