Bruno Appolloni


We discuss unprecedented, albeit rudimentary, tools to infer the evolution of a point process where the available samples are both truncated and non independently drawn. To achieve this goal, we lay in an intermediate domain between probability models and fuzzy sets, still maintaining probabilistic features of the employed statistics as the reference KPI of the tools. The overall strategy is to frame the problem within the Algorithmic Inference framework and use a sort of kernel trick to distort the seeds of the observed variable so as to render them an iid sample of a random variable in a proper feature space. Numerical results highlight the suitability of the proposed tools.

Included in

Probability Commons