R中的修正指数分布

Question

I would like to make my own probability density function in R to simulate some things from a paper.

It is somehow similar to exponential distribution but what i really want to do is to redefine the exponential distribution into a "modified" one...

Is there such a way to do this? Thanks.

---

I want to simulate this:

b(x) = {(µ/p)(e ^ (-(µx-q)/p) , x > q(xbar) and 0 otherwise }

xbar is an x with a line above it, mean, average

Answer 1

您是否知道从下面截断的指数分布仍然是指数分布？

Answer 2

I believe you can do what you want with ?sample . Use your known distribution function b(x) to generate a vector of probabilities, say bprob , then

sample(x,size, replace=TRUE,prob=bprob)

There are some very interesting methods for generating samples from arbitrary distributions. See for example, Normal RandomNumbers:UsingMachine AnalysisTo Choosethe BestAlgorithm WH PAYNE WashlngtonState University (somewhere on the web), or Numerical Recipes in C The Art of Scientific Computing Second Edition William H. Press,Saul A. Teukolsky et al, chapter 7.3