Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Amy <dartmouthemails@yahoo.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: kdensity with few (/aggregated) data points |
Date | Wed, 30 Jun 2010 08:23:11 -0700 (PDT) |
Hi, I just thought to re-phrase my question. I've noticed that if I have very few data points (e.g. 10) then kdensity gives me something jagged even if I specify a Gaussian kernel (regardless of the bandwidth). If the reason I have so few data points is because I have aggregate data, e.g. data for each decile of a population, is there any way to make this smoother? Why is it that histogram X, bin(10) kdensity kdenopts(gauss) will give me something that looks smoother? Thank you. --- On Sat, 6/26/10, Amy <dartmouthemails@yahoo.com> wrote: > From: Amy <dartmouthemails@yahoo.com> > Subject: kdensity with binned data > To: statalist@hsphsun2.harvard.edu > Date: Saturday, June 26, 2010, 3:45 AM > Hi, > > I have aggregate data for each decile of a population. When > I try to > plot the kernel density estimator (for sake of reference, > you can define > variable "test" to be 10, 20, 40, 50, 60, 70, 80, 90, 95, > 99 and then > try > > kdensity test, kernel(gauss) > > with various bandwidths), its peaks at each of the values > are pointed > rather than smooth. Yet if I try > > histogram test, bin(10) kdensity kdenopts(gauss) > > the density will appear much smoother, even for the same > bandwidth and > kernel. Comparing the results with R's "density(test, > kernel=c("gaussian"), weights=NULL, window=kernel, n=100)", > R's density > also looks smooth (Stata's "kdensity test, kernel(gauss)" > won't let me > specify n(100) since the number of binned datapoints I have > is 10). > Ultimately I would like to use those 10 binned data points > to plot the > density across the whole population and sample from it so > that I can > infer some values for the person at the 26th percentile, > the 27th > percentile, etc. > > Can Stata handle binned data like this? > > Thank you. > > > > * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/