DNAprofiles (version 0.3.1)

dists.product.pair: Distribution of product of several discrete random variables as product X*Y

Description

Distribution of product of several discrete random variables as product X*Y

Usage

dists.product.pair(dists, n.max = 1e+06, appr = FALSE, appr.method = 1L, n.max.appr = 1000, r0 = 0.01, R = 1.05)

Arguments

dists

a list of distributions

n.max

maximum number of mass points of discrete distribution used in the process

appr

if TRUE, then the distributions are shrunken (approximated), if necessary, to not exceed n.max

appr.method

integer: 1 (merge mass points to lower bound); 2 (merge to upper bound)

n.max.appr

maximum number of mass points of shrunken distributions

r0

numeric, relative tolerance used in first step of shrinking the distributions

R

numeric, r0 is multiplied with R until the number of mass points is at most n.max.appr

Value

list with named sublists:

• cumdist1: a list with vectors x, Fx
• dist2: a list with vectors x, fx

Examples

data(freqsNLngm)

set.seed(123)
x <- sample.profiles(1,freqsNLngm)

# per locus distribution of kinship index
dists <- ki.dist(x,hyp.1="FS",hyp.2="UN",hyp.true="UN")

n <- sapply(dists,function(x) length(x$fx))
prod(n) # too many outcomes to store!
# but, for two subsets of the loci, the distribution can be obtained
pair <- dists.product.pair(dists)
str(pair) # with these, we can compute exceedance probabilities quickly

# obtain the cdf as a function
cdf <- dist.pair.cdf(pair)
cdf(1)

# plot the cdf
x0 <- seq(from=-10,to=5,length=50)
plot(x0,cdf(10^x0),type="l",xlab="x",ylab="Fn(x)")