NEURAL NETWORK PERCEPTRON ELEMENTAL MODEL FOR PETROPHYSICAL APPLICATIONS
This example is a logic diagram of a GeoNeurale concept.
It illustrates the workflow of a Neural Network for the estimation
process of facies logs from petrophysical properties.
The training consists in updating the weights until
a weights zero change is reached.
This weight calculation is normally performed applying
logical lerning rules.
Here the Delta lerning rule of Widrow/Hoff or the Hebb lerning
rule can be applied.
Schematic two layers ( no hidden layers ) model example
i1 , i2 : input values 1st and 2nd Units
w : applied weights
o : operators
r : partial results
t : threshold function