runi(n), n a positive integer |

runi(count) generates a vector of count pseudo-random uniforms on the interval 0 to 1. If the random number generator has not been initialized by setseeds(), setoptions() or previous use of rbin(), rnorm(), rpoi() or runi(), the generator's "seeds" will be initialized automatically using the current time and date, and their values will be printed out. You can generate uniform random variables on the interval (a,b), a < b by Cmd> x <- a + (b - a)*runi(n) You can generate the discrete uniform distribution on the integers 1, 2, ..., m by Cmd> x <- ceiling(m*runi(n)) This is helpful when sampling with replacement from the rows of a data vector of matrix. When Q() is a macro or function computing the quantile function (inverse cumulative distribution function) of a continuous random variable (invnor() or invchi()) for example), Q(runi(n) [,parameters]) generates a random sample from that distribution. Cmd> invstu(runi(5),3) # Student's t on 3 d.f. (1) 0.43734 0.34297 0.054439 -0.0017229 -0.32894 Cmd> invF(runi(5),5,30) # F on 5 and 30 d.f. (1) 0.45207 2.2247 0.52716 0.29218 1.506 The functions that you can use directly this way are invbeta(), invchi(), invF(), invgamma(), invnor(), and invstu(). In principle you could also use invdunnett() and invstudrng(), but that is not advisable because they are so computationally intensive. See also topics setseeds(), getseeds(), setoptions(), rnorm(), rbin(), rpoi(), subtopic 'options:"seeds"'

Gary Oehlert 2003-01-15