Answers are grounded in this paper and cite their sources.
It is shown how a very simple distribution, the Bernoulli random variable, exhibits all possible normalised kurtosis values when varying the probability of its two events. This result means that signals of any normalised kurtosis may be modelled by this distributiori, which is of relevance within the framework of blind source separation based on fourth-order cumulants.
No takes yet. Share an insight, caveat, or question.
Zarzoso et al. (1998) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: