7

I have a problem where I want to generate a set of random integer values between 1 and 5 inclusive using a probability distribution.

Poisson and Inverse Gamma are two distributions that show the characteristics I am after (majority at mean, less higher numbers) that I have found.

I am looking at using Apache Commons Math but I wasn't sure how to generate the numbers I wanted using the distributions available.

jmc
  • 620
  • 14
  • 24
  • 1
    Well, here's a starting point: http://commons.apache.org/proper/commons-math/javadocs/api-3.2/org/apache/commons/math3/distribution/PoissonDistribution.html – Yuushi May 08 '13 at 08:13

2 Answers2

12

From your problem description, it sounds like you actually want a sample generated from a discrete probability distribution, and you can use EnumeratedIntegerDistribution for this purpose. Choose appropriate probabilities for each of your integers, maybe something like the following would meet your needs:

int[] numsToGenerate           = new int[]    { 1,   2,    3,   4,    5   };
double[] discreteProbabilities = new double[] { 0.1, 0.25, 0.3, 0.25, 0.1 };

EnumeratedIntegerDistribution distribution = 
    new EnumeratedIntegerDistribution(numsToGenerate, discreteProbabilities);

int numSamples = 100;
int[] samples = distribution.sample(numSamples);

Just tweak the discreteProbabilities values to whatever you require.

iainmcgin
  • 2,691
  • 1
  • 18
  • 24
0

I would just produce uniformly distributed random numbers then pass them into the distribution function you want, so if the distribution function was x^2

import java.util.ArrayList;

import java.util.Random;



    public class Test{

        public static void main(String args[]){
            Test t=new Test();
        }

        public Test(){

            Random rnd=new Random();

            ArrayList<Double> data=new ArrayList<Double>();

            for(int i=0;i<1000;i++){
                data.add(useFunction(rnd.nextDouble()));
            }

        }

        public double useFunction(double in){
            return in*in;
        }
    }
Richard Tingle
  • 16,906
  • 5
  • 52
  • 77