Connectionist Forecasting Of Earthquakes: Method And Preliminary Results
Price
Free (open access)
Volume
19
Pages
9
Published
1997
Size
48 kb
Paper DOI
10.2495/AI970181
Copyright
WIT Press
Author(s)
E. Ivo Alves
Abstract
. This method is based on the integration by an artificial neural network of forecasting tools that were originally developed for financial analysis, the so-called financial oscillators. A preliminary test is performed on seismic data from the Azores, and that test shows a good predicting ability, better than would be obtained by classical probabilistic methods. 1 Introduction. Although Man can alter the environment, and sometimes drastically does so, some of the most severe environmental changes arise from natural processes. These processes, given their magnitudes are, for the time being, beyond any possibility of control. The only hope for protection from natural catastrophes - volcanic eruptions, floods, droughts, earthquakes - is in their prediction. Earthquake prediction has taken different aspects:
Keywords