Hi Mike, if you can decompose the bimodal distribution into (eg two) known forms, then you could try a stepwise approach, eg:
If uniform < 0.5 then double it and use it to draw from the inverse cdf of A, else, double (uniform - 0.5) and use it to draw from the inverse cdf of B. You can change the cutoff and the weights to suit your need. It's best to double-check by plotting an empirical density of the numbers generated. I hope that this helps, Andrew On Thu, May 29, 2008 at 04:05:29PM -0600, Mike Williams wrote: > Hello R Users, > > I am doing a Latin Hypercube type simulation. I have found the > improvedLHS function and have used it to generate a bunch of properly > distributed uniform probabilities. Now I am using functions like qlnorm > to transform that into the appropriately lognormal or triangularly > distributed parameters for my modes. However I have a parameter which I > believe is bimodally distributed, could someone please point me at an > appropriate function equivalent to qlnorm which I can use, because for > some reason I have been unable to find one. It occurs to me that maybe > one doesn't exist, in which case could someone give me some other idea > of how to accomplish this goal? > > Thanks, > > Mike > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. -- Andrew Robinson Department of Mathematics and Statistics Tel: +61-3-8344-6410 University of Melbourne, VIC 3010 Australia Fax: +61-3-8344-4599 http://www.ms.unimelb.edu.au/~andrewpr http://blogs.mbs.edu/fishing-in-the-bay/ ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
