diveMove (version 1.6.2)

rmixexp: Generate samples from a mixture of exponential distributions

Description

rmixexp uses a special definition for the probabilities p_i to generate random samples from a mixed Poisson distribution with known parameters for each process. In the two-process case, p represents the proportion of "fast" to "slow" events in the mixture. In the three-process case, p_0 represents the proportion of "fast" to "slow" events, and p_1 represents the proportion of "slow" to "slow" *and* "very slow" events.

Usage

rmixexp(n, p, lambdas)

Value

vector of samples.

Arguments

n

integer output sample size.

p

numeric probabilities for processes generating the output mixture sample.

lambdas

numeric lambda (rate) for each process.

Examples

## Draw samples from a mixture where the first process occurs with
## p < 0.7, and the second process occurs with the remaining
## probability.
p <- 0.7
lda <- c(0.05, 0.005)
(rndprocs2 <- rmixexp(1000, p, lda))

## 3-process
p_f <- 0.6    # fast to slow
p_svs <- 0.7  # prop of slow to (slow + very slow) procs
p_true <- c(p_f, p_svs)
lda_true <- c(0.05, 0.01, 8e-4)
(rndprocs3 <- rmixexp(1000, p_true, lda_true))