Skip to content
Deepbox
21intermediate

Random Sampling & Distributions

Draw random numbers from common distributions. setSeed makes every later draw reproducible, which helps with simulations, tests and repeatable experiments.

What you will learn

  • Use deepbox/random for setSeed, rand, randn, randint, uniform, normal, binomial, poisson, exponential, gamma, beta, multivariateNormal, choice, permutation, Generator.
  • Use deepbox/ndarray for tensor.
  • Draw random numbers from common distributions. setSeed makes every later draw reproducible, which helps with simulations, tests and repeatable experiments.

Source

index.ts
/** * Example 21: Random Sampling & Distributions * * Draw random numbers from common probability distributions. Calling * setSeed() makes every later draw reproducible, which is what you want for * simulations, tests and repeatable experiments. */ import { tensor } from "deepbox/ndarray";import {  beta,  binomial,  choice,  exponential,  Generator,  gamma,  multivariateNormal,  normal,  permutation,  poisson,  rand,  randint,  randn,  setSeed,  uniform,} from "deepbox/random"; console.log("=== Random Sampling & Distributions ===\n"); // Set the seed for reproducibilitysetSeed(42);console.log("Random seed set to 42 for reproducibility\n"); // Seeding again with the same value repeats the same numbers.const firstRun = rand([3]);setSeed(42);const secondRun = rand([3]);console.log(`Same seed, same draws: ${Number(firstRun.eq(secondRun).all().item()) === 1}\n`); // Uniform distribution [0, 1)console.log("1. Uniform Distribution [0, 1):");const uniformSamples = rand([5]);console.log(`${uniformSamples.toString()}\n`); // Standard normal distributionconsole.log("2. Standard Normal Distribution (mean=0, std=1):");const normalSamples = randn([5]);console.log(`${normalSamples.toString()}\n`); // Random integersconsole.log("3. Random Integers [0, 10):");const intSamples = randint(0, 10, [8]);console.log(`${intSamples.toString()}\n`); // Custom uniform distributionconsole.log("4. Uniform Distribution [-5, 5]:");const customUniform = uniform(-5, 5, [6]);console.log(`${customUniform.toString()}\n`); // Custom normal distributionconsole.log("5. Normal Distribution (mean=100, std=15):");const customNormal = normal(100, 15, [6]);console.log(`${customNormal.toString()}\n`); // Binomial distribution (coin flips)console.log("6. Binomial Distribution (n=10, p=0.5):");const binomialSamples = binomial(10, 0.5, [8]);console.log(binomialSamples.toString());console.log("(Number of heads in 10 coin flips)\n"); // Poisson distributionconsole.log("7. Poisson Distribution (lambda = 3):");const poissonSamples = poisson(3, [8]);console.log(poissonSamples.toString());console.log("(Number of events with rate lambda = 3)\n"); // Exponential distributionconsole.log("8. Exponential Distribution (scale=2):");const expSamples = exponential(2, [6]);console.log(expSamples.toString());console.log("(Time between events)\n"); // Gamma distributionconsole.log("9. Gamma Distribution (shape=2, scale=2):");const gammaSamples = gamma(2, 2, [6]);console.log(`${gammaSamples.toString()}\n`); // Beta distributionconsole.log("10. Beta Distribution (alpha = 2, beta = 5):");const betaSamples = beta(2, 5, [6]);console.log(betaSamples.toString());console.log("(Values between 0 and 1)\n"); // Correlated samples from a multivariate normal. Each row is one draw.console.log("11. Multivariate Normal (mean [0, 0], covariance [[1, 0.8], [0.8, 1]]):");const mvn = multivariateNormal(  tensor([0, 0]),  tensor([    [1, 0.8],    [0.8, 1],  ]),  3);console.log(`${mvn.toString()}\n`); // Sample from a list, with and without replacement, and shuffle.console.log("12. Choice and Permutation:");const items = tensor([10, 20, 30, 40, 50]);console.log(`With replacement:    ${choice(items, 4).toString()}`);console.log(`Without replacement: ${choice(items, 4, false).toString()}`);console.log(`Shuffled copy:       ${permutation(items).toString()}`);console.log(`Shuffled 0..4:       ${permutation(5).toString()}\n`); // A Generator has its own state, so it does not disturb the global seed.console.log("13. Independent Generator:");const rngA = new Generator(7);const rngB = new Generator(7);console.log(`Two generators with seed 7 agree: ${rngA.random() === rngB.random()}`);

Output

Console output
Console output only. The numbers are the same on every run because of the seed. They differ from 1.0.0 because the random streams changed in 1.5.0.