Ideas on how to show that the variance of a doubly-stochastic Poisson process(aka a Cox process) driven by a homogeneous(stationary) Poisson process of intensity $\lambda$ is $3\lambda$ ?

I've come to this conjecture from simulation with the following Java code

package stochastic.processes;

import java.util.function.DoubleSupplier;

import org.apache.commons.math3.distribution.ExponentialDistribution;
import org.apache.commons.math3.random.RandomGenerator;
import org.apache.commons.math3.random.Well19937c;

public class CoxProcess implements DoubleSupplier {

    private final ExponentialDistribution intensityDistribution;

    RandomGenerator generator;

    public CoxProcess( double intensity )
        this( new Well19937c(), intensity );

    public CoxProcess( RandomGenerator generator, double intensity )
        this.generator = generator;
        intensityDistribution = new ExponentialDistribution( 1.0 / intensity);

    public double getAsDouble()
        ExponentialDistribution dist = new ExponentialDistribution(generator, intensityDistribution.sample());
        return dist.sample();

enter image description here



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.