Search

We describe the main structure and outcomes of the new probabilistic seismic hazard model for Italy, MPS19 [Modello di Pericolosità Sismica, 2019]. Besides to outline the probabilistic framework adopted, the multitude of new data that have been made available after the preparation of the previous MPS04, and the set of earthquake rate and ground motion models used, we give particular emphasis to the main novelties of the modeling and the MPS19 outcomes. Specifically, we (i) introduce a novel approach to estimate and to visualize the epistemic uncertainty over the whole country; (ii) assign weights to each model components (earthquake rate and ground motion models) according to a quantitative testing phase and structured experts’ elicitation sessions; (iii) test (retrospectively) the MPS19 outcomes with the horizontal peak ground acceleration observed in the last decades, and the macroseismic intensities of the last centuries; (iv) introduce a pioneering approach to build MPS19_cluster, which accounts for the effect of earthquakes that have been removed by declustering. Finally, to make the interpretation of MPS19 outcomes easier for a wide range of possible stakeholders, we represent the final result also in terms of probability to exceed 0.15 g in 50 years., peer-reviewed

A procedure is presented to evaluate the statistical significance of the time association of two different series of events, which are represented by stationary, single valued, stochastic point processes., Statistica; Vol 52, No 1 (1992)

We developed a procedure to find change-points which seems both widely applicable and easy to implement on computer. The number of regimes is not needed a priori. Monte Carlo simulations showed that it is efficient also at low sample size. The approach is fully nonparametric and can be applied without knowing the statistical distribution of each subset. Confidence intervals are evaluated by a Monte Carlo simulation and this requires the latter distributions to be given explicitly. These are determined by Kolumogorov-Smirnov one sample goodness-of-fit test applied to all the different sets identified in the change-point analysis. The procedure is presented with a practical application to Gaussian and Poisson series., Statistica; Vol 46, No 1 (1986)

L'articolo affronta il problema della stima del parametro p della distribuzione geometrica e dei relativi intervalli di confidenza. L'argomento è di notevole importanza in sismologia perchè si è trovato che la magnitudo dei terremoti segue una distribuzione esponenziale (la ben nota legge di Gutenberg-Richter) e perciò, a seguito del raggruppamento dei dati in classi di ampiezza finita, ci si attende che i campioni sperimentali si adattino alla distribuzione geometrica. Per motivi di completezza, diamo un rapido cenno dei metodi per la stima "puntuale" del parametro p. Obiettivo principale dell'articolo è tuttavia la determinazione degli intervalli di confidenza relativi a p. Un'analisi accurata ci consente di discutere e confrontare soluzioni esatte e soluzioni asintotiche. Sotto forma di tabelle vengono presentati gli intervalli di confidenza calcolati per i valori standard del coefficiente di confidenza. Le tabelle sono intese come un utile strumento di lavoro, in quanto bastano gli elementari procedimenti algebrici di interpolazione e di estrapolazione per calcolare anche ogni soluzione non esplicitamente compresa nelle tavole., Statistica; Vol 46, No 2 (1986)

Computational pattern recognition is an invaluable tool in understanding the phenomenology of complex processes and represents the first step towards their effective physical modelling. So far its practical applications have been rather limited, but its potential appears great in all the fields characterized by a behaviour complex enough to defy intuition. We discuss in detail pattern recognition algorithms of the "logic" type, which proved effective in a number of diverse fields., Statistica; Vol 52, No 2 (1992)