A simple model for fractional non-Gaussian processes

Fractionally differenced Gaussian noise processes have been found to be useful for modeling the long-term dependencies exhibited by many real processes. Non-Gaussian processes have also been observed which exhibit long-term dependencies, but simple models have been lacking. A simple model which is a generalization of the fractionally differenced Gaussian noise process is proposed. The model should be useful for generating the random vectors needed for simulating systems which produce some result based on a finite length block of observations