3

I know this must be pretty basic, but what is the proper, accurate way to plot the PDF of some sample data that you know comes from some pop. distribution, like if you generated it using rnorm() or rexp()?

The reason I ask is because I know a lot of people use density(), and then input that into plot(), but the density() function seems too arbitrary to be accurate; for example, it is inaccurate when it approximates negative value for data that came from the exponential distribution, which does not possess neg. values.

So could someone recommend me a more fine-tuned method to accomplish plotting sample PDFs?

halfer
  • 19,824
  • 17
  • 99
  • 186
Coolio2654
  • 1,589
  • 3
  • 21
  • 46
  • 1
    A similar question (or at least related), was asked on [stats.se] - my [answer](https://stats.stackexchange.com/a/71291/1390) covers the bit about the implausible negative values for a variable. – Gavin Simpson Sep 12 '17 at 19:50

2 Answers2

1

The density function performs kernel density estimation (KDE). To find the best KDE for your dataset, you should tune the bandwidth (parameter bw). Here's a paper that discusses KDE and bandwidth selection: http://www.stat.washington.edu/courses/stat527/s13/readings/Sheather_StatSci_2004.pdf

Or for a simpler approach, you can try out different bandwidth methods to pass to bw: https://stat.ethz.ch/R-manual/R-devel/library/stats/html/bandwidth.html

The current default, "nrd0", is there for historical reasons. I find "ucv" and "bcv" have worked better for my datasets.

Kelli-Jean
  • 1,417
  • 11
  • 17
  • 1
    These will still leak density into negative values if you have data values close to 0, unless you have a very small bandwidth indeed. – Gavin Simpson Sep 12 '17 at 19:52
0

ggplot does help take care of negative values when they are not appropriate. It can be used in the following manner:

ggplot(df, 
       aes(x=contVar, fill = "green")) + 
  geom_density(alpha=.3)

I would also take a look at this post in cross validated

DataTx
  • 1,839
  • 3
  • 26
  • 49
  • While I thought the above answer addressed my question in a more technical and thorough manner, your link is a great resource and want to thank you for it. – Coolio2654 Sep 13 '17 at 20:19