A neural network is a highly connected system, whose elementary unit is the artificial neuron that reproduces in some way the nonlinear function of a biological neuron of the brain. Neuron has many input lines (dendrites) and only one output line (axon), which, by means of a connection (synapses), goes to another neuron or is taken as one output of the system. Using this approach we have developed a neural system to estimate the magnitude of forthcoming earthquakes. The neural system we use for the earthquake magnitude estimation problem is a multilayer backpropagation feedforward neural network. This is usually referred to as MLP (MultiLayer Perceptron). We have validated this estimation scheme to the sequence of earthquake magnitudes of some seismic areas of Italy. Our results reveal a good agrrement between the estimated with the observed magnitudes.
A Neural Network approach to estimate the magnitude of forthcoming earthquakes
Telesca L;Viggiano M
2004
Abstract
A neural network is a highly connected system, whose elementary unit is the artificial neuron that reproduces in some way the nonlinear function of a biological neuron of the brain. Neuron has many input lines (dendrites) and only one output line (axon), which, by means of a connection (synapses), goes to another neuron or is taken as one output of the system. Using this approach we have developed a neural system to estimate the magnitude of forthcoming earthquakes. The neural system we use for the earthquake magnitude estimation problem is a multilayer backpropagation feedforward neural network. This is usually referred to as MLP (MultiLayer Perceptron). We have validated this estimation scheme to the sequence of earthquake magnitudes of some seismic areas of Italy. Our results reveal a good agrrement between the estimated with the observed magnitudes.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
