0
$\begingroup$

Suppose I have a discrete stochastic process, in the form of $$x_{t+1} = x_t + \varepsilon_t$$ where $\varepsilon_t$ is the random noise. The caveat is by examining the existing data, $\varepsilon_t$ does not follow a normal distribution. Suppose it follows a Cauchy distribution or more well-behavior distributions such as hyperbolic distribution. Is there any general procedure to convert this discrete stochastic process into a continuous stochastic process described by $$dx = \mu(x, t) \, dt + \sigma(x, t) \, dW$$ where $\mu$ is the drift, $\sigma$ is the square root of variance of the noise, and $dW$ is Wiener process?

$\endgroup$
0

1 Answer 1

1
$\begingroup$

You would first need to check whether Donsker's applies to your process https://math.stackexchange.com/questions/274685/joint-convergence-and-donskers-theorem. If not, it could be that the process has some different limiting process $L$ eg. some Levy process https://en.wikipedia.org/wiki/Cauchy_process and so you will need to study the stochastic integral wrt to that process.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.