RandSamples()

Summary

Return an array of random samples of the given type of random distribution and parameters defined in the options array.

 

Changes

Added to Version 2019.

 

Syntax

sample_vector = RandSamples(int n, string type, opt array params)

 

Argument	Contents
n	The number of samples to return (in an array)
type	One of the following strings:

 

Type	Description
Uniform	Random numbers between 0.0 (inclusive) and 1.0 (exclusive) with uniform distribution. If option "population" is provided, then the return values will be items in the array rather than values between 0.0 and 1.0.
Discrete	Random numbers between 1 and n based on the probability densities defined by option "weight", where n is the length of the "weight" array. An option "population", which has the same length as "weight", can also be given if you want the return values to be drawn from the population array instead of the 1-based index to the population bins. Items in the population array can be int, string, double, or color.
Binary	Random numbers of 1 and 0, with probability of 1 given by optional parameter "p" (default is 0.5).
NegativeExp	Random numbers of negative exponential distribution. Requires parameter "lambda"
Poisson	Random integers from the Poisson distribution. Requires parameter "lambda".
Norm	Random numbers of Normal distribution. Optional parameters are mean "mu" (0.0) and standard derivation "sigma" (1.0).
LogNorm	Random numbers of Log Normal distribution. Optional parameters are mean "mu" (0.0) and standard derivation "sigma" (1.0).
Gumbel	Random numbers of Gumbal distribution. Optional parameters are mode "mu" and standard derivation "sigma".
Gamma	Random numbers of Gamma distribution. Parameters are "shape" (k) and "scale" (theta).

 

The options supported vary by distribution type:

 

Uniform:

Option	Type	Contents
population	array or vector	If provided, returned values will be drawn from this array or vector. If not provided, returns a value between 0 and 1.

 

Discrete:

Option	Type	Contents
population	array or vector	If provided, returned values will be drawn from this array or vector. If not provided, returns an integer between 1 and the length of the Weight array or vector.
weight	array or vector	Defines the weights to use for sampling. Defaults to an array or vector of size n filled with the value 1.

 

Binary:

Option	Type	Contents
p	real or array	A value between 0 and 1.0 that describes the probability of sampling a 1. Defaults to 0.5.

 

NegativeExp and Poisson:

Option	Type	Contents
lambda	real or array	Describes the shape of the distribution. Must be greater than 0. Defaults to 1.

 

Norm, LogNorm, and Gumbell:

Option	Type	Contents
mu	real or array	The mean of the distribution for "Norm" and "LogNorm". For "Gumbell", it is the mode. Defaults to 0.
sigma	real or array	The standard deviation of the distribution. Defaults to 1.

 

Gamma:

Option	Type	Contents
shape	real or array	Controls the shape of the function. Defaults to 1.
scale	real or array	Controls the scale of the function. Defaults to 1.

 

Returns

A vector of random samples.

 

Notes

• The options array is an optional argument.

 

Examples

// Example 1

SetRandomSeed(42)

params = null

params.population = { "A", "B", "C", "D", "E" }

samples = RandSamples(10, "Uniform", params)

// samples = { "vector", { "A", "B", "A", "A", "A", "C", "C", "A", "A", "D" } }

 

// Example 2

SetRandomSeed(42)

params = null

params.population = { "A", "B", "C", "D", "E" }

params.weight = { 20, 45, 18, 4, 22 }

samples = RandSamples(10, "Discrete", params)

// samples = { "vector", { "A", "B", "A", "A", "A", "B", "B", "B", "A", "C" } }

 

// Example 3

SetRandomSeed(42)

params = null

params.mu = 1

params.sigma = 2

samples = RandSamples(5, "Norm", params)

// samples = { "vector", { -0.487258, -0.486241, -4.497264, -1.318424, -1.180147 } }

 

// Example 4

SetRandomSeed(1)

params = null

params.p = .75

samples = RandSamples(10, "Binary", params)

// samples = { "vector", { 1, 1, 0, 1, 1, 1, 1, 0, 0, 1 } }

 

// Example 5 - using vectors/arrays for parameters

SetRandomSeed(42)

params = null

params.mu = {1.2, 4.6, 17.2}

// Vectors work, too:

// params.mu = A2V({1.2, 4.6, 17.2})

samples = RandSamples(3, "Norm", params)

// samples = { "vector", { 0.456371, 3.856879, 14.451368 } }

 

See Also

Function	Summary
SetRandomSeed()	Seeds the RandomNumber() function