Lecture Notes in Computer Science Volume 2859 In this paper we analyze seismic signals recorded in September 1997 in Stromboli (Sicily) during explosive eruptions. First, we analyze the data via an unsupervised techniques using the Mixture of Gaussians algorithm (MoG) and the Principal Component Analysis (PCA) to discover the structure of the data. Experts distinguish two types of signals related to two different type of Strombolian explosive eruptions (Type 1 and Type 2). Using the MoG algorithm we can distinguish two classes that, with a good agreement, correspond to the two types of explosions given by experts. As a second step, we implement an supervised automatic system in order to discriminate between the two different types of explosive eruptions. The automatic system based on the MLP achieve a correct classification percentage of more then 98% on the test set (and 100% on the training).

Supervised and Unsupervised analysis applied to Strombolian E.Q.

MARINARO, Maria;SCARPETTA, Silvia
2003

Abstract

Lecture Notes in Computer Science Volume 2859 In this paper we analyze seismic signals recorded in September 1997 in Stromboli (Sicily) during explosive eruptions. First, we analyze the data via an unsupervised techniques using the Mixture of Gaussians algorithm (MoG) and the Principal Component Analysis (PCA) to discover the structure of the data. Experts distinguish two types of signals related to two different type of Strombolian explosive eruptions (Type 1 and Type 2). Using the MoG algorithm we can distinguish two classes that, with a good agreement, correspond to the two types of explosions given by experts. As a second step, we implement an supervised automatic system in order to discriminate between the two different types of explosive eruptions. The automatic system based on the MLP achieve a correct classification percentage of more then 98% on the test set (and 100% on the training).
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11386/1063371
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? 3
social impact