In this paper we propose a clustering technique, based on the maximization of the likelihood function defined from the generalization of a model for seismic activity (ETAS model, (Ogata (1988))), iteratively changing the partitioning of the events. In this context it is useful to apply models requiring the distinction between independent events (i.e. the background seismicity) and strongly correlated ones. This technique develops nonparametric estimation methods of the point process intensity function. To evaluate the goodness of fit of the model, from which the clustering method is implemented, residuals process analysis is used.
|Titolo della pubblicazione ospite||Data Analysis, Classification and the Forward Search|
|Numero di pagine||8|
|Stato di pubblicazione||Published - 2006|
|Nome||STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION|