Random values, as TDA's rd()/rdn()

The R-side equivalent of what nvar()'s X = rd(0, 3) or Y = sin(X) + rd would generate inside TDA: tda_rd(n, a, b) is uniform on [a, b] (rd's two-argument form, a + random1() * (b - a)), and tda_rd(n) alone – rd with no arguments – is uniform on [0, 1], not the wider, arbitrary range a hand-rolled substitute might reach for. tda_rdn is the same idea for rdn, a standard normal.

Usage

tda_rd(n, a = 0, b = 1)

tda_rdn(n, mean = 0, sd = 1)

Arguments

n

how many values to draw.

a, b

the range, for tda_rd; TDA's default (no arguments) is [0, 1].

mean, sd

the normal distribution's parameters, for tda_rdn; TDA's rdn is always standard (0, 1).

Value

A numeric vector of length n.

Details

These use R's random number generator, not tda_runif's reproduction of TDA's exact stream – for building a data frame to pass to TDA (rather than reproducing an existing TDA example's numbers bit for bit), R's generator is what set.seed() already controls.

Examples

set.seed(1)
x <- tda_rd(200, 0, 3)      # nvar()'s X = rd(0, 3)
y <- sin(x) + tda_rd(200)   # nvar()'s Y = sin(X) + rd
round(c(mean = mean(x), sd = sd(x)), 2)   # near 0 and 3
#> mean   sd 
#> 1.51 0.90