Bivariate Gaussian Generator

[grabonlee]

Hi

Please take a look at the code below. I'm trying to generate an array of random numbers with mean(1.0,1.0) and covariance(1.0, 0.0,0.0,1.0).

I'm stuck with how to make the number of samples(random numbers) up 10000.


import java.util.Rando m;

public class BivariateGaussi anGenerator
{

static float mu1,mu2,co1,co2 ,co3,co4,covari ance1,
covariance2,cov ariance3,covari ance4,mean1,mea n2;



public static void setMean(float mu1,float mu2 )
{
mean1=mu1;
mean2=mu2;
}

public void setCovariance(f loat co1,float co2,float co3,float co4)
{
covariance1=co1 ;
covariance2=co2 ;
covariance3=co3 ;
covariance4=co4 ;
}

public static float[] generateRandom( )
{
int count = 100000;

Random generate = new Random();
float[] genNumbers = new float[2];


for(int i=0;i<count;i++ )
{
genNumbers[i] = Math.abs((float )generate.nextG aussian());
}
return genNumbers;
}

public static void main(String[] args)

{

BivariateGaussi anGenerator bgg = new BivariateGaussi anGenerator();

bgg.setMean(1.0 f,1.0f);
bgg.setCovarian ce(1.0f,0.0f,0. 0f,1.0f);

float[] x = bgg.generateRan dom();


System.out.prin tln("["+x[0]+" , " +x[1]+"]");

}

}

[JosAH]

You have your indexes all messed up, i.e. you define an array that can hold
two floating point numbers and you try to stick 100,000 N(0,1) numbers in it;
that will throws an ArrayIndexOutOf BoundsException .

kind regards,

Jos

[grabonlee]

You have your indexes all messed up, i.e. you define an array that can hold
two floating point numbers and you try to stick 100,000 N(0,1) numbers in it;
that will throws an ArrayIndexOutOf BoundsException .

kind regards,

Jos

Thanks

Yes I want to create array.length = 2. I also want each array to generate up 100,000 random numbers, which would be bound by a defined mean and covariance.
I want to also plot a graph. Like in Matlab, the more the random samples, the clearer the graph. What can I do?

[JosAH]

Thanks

Yes I want to create array.length = 2. I also want each array to generate up 100,000 random numbers, which would be bound by a defined mean and covariance.
I want to also plot a graph. Like in Matlab, the more the random samples, the clearer the graph. What can I do?

Didn't that method I copied from Abramowitz and Stegun help you? See your
other thread for that.

kind regards,

Jos