The project is aimed at investigating advanced Pattern Recognition techniques for the analysis of seismic signals, coming from different volcanoes, mainly employing probabilistic techniques such as Hidden Markov Models.
The project is carried out in collaboration with Mauricio Orozco Alzate, associate professor of the Universidad Nacional de Colombia, sede Manizales. Such project is aimed at exploiting two important facts: from one hand, M. Bicego has gained a relevant knowledge and expertise on HMM (both from the theoretical and the applicative point of view), being very active in the field since 2001. On the other hand, the Pattern Recognition and Signal Processing group of the University of Colombia (Manizales) has one of the largest database in the world of seismic signals, which contains more than 16 years of daily recordings, taken from at least 5 stations and relative to 4 volcanoes. They have been developing Pattern Recognition systems based on these data since 2005, gaining an invaluable context-related expertise. The analysis of such peculiar database is a real challenge also from the theoretical point of view, but hardly leading to sensible results without face to face interaction with the experts in the field.