3

I have plotted this curve using the default kernel density estimate function in R I have plotted this curve using the default kernel density estimate function in R.

I am looking for some standard distribution which could be close to this. Is there any standard bimodal distribution ?

the_dude
  • 35
  • 1
  • 5
  • 1
    The short answer is no (although some readers might be tempted to suggest a two-component Gaussian mixture model, which arguably is a "standard bimodal distribution"). A more relevant issue is *why are you looking for such a distribution*? What is the purpose of your analysis? If you could include such information in edits to this post you might obtain answers that ultimately are more useful. – whuber May 27 '15 at 18:29

1 Answers1

4

While I am not aware of anything that can be called ''standard'' bimodal distribution, in this particular case, mixture normal distribution seems to be appropriate at first glance. The pdf of such distribution is essentially the linear combination of two (or more) - not necessarily equal means or equal variances - normal distribution's pdf. (Thus the mixing weight is also a further parameter.)

R package mixtools provides tools for estimating such distributions.

Tamas Ferenci
  • 3,143
  • 16
  • 26