Created
November 20, 2012 13:20
-
-
Save robertovalenzuela91/4117900 to your computer and use it in GitHub Desktop.
Perceptron
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import random | |
class Neurona: | |
def __init__(self, nEntradas, alfa): | |
self.nEntradas = nEntradas | |
self.alfa = alfa | |
self.pesos = [] | |
for i in range(nEntradas): | |
self.pesos.append(random.random()) | |
self.pesos.append(random.random()/2.0)# Umbral | |
def sumador(self, entradas): | |
suma = 0.0 | |
for i in range(self.nEntradas): | |
suma += self.pesos[i]*entradas[i] | |
return suma | |
def funcion_no_lineal(self, entradas): | |
suma = self.sumador(entradas) | |
suma -= self.pesos[-1] | |
if suma >= 0: | |
return 1 | |
else: | |
return 0 | |
def entrenar(self, entradas, esperado): | |
salida = self.funcion_no_lineal(entradas) | |
entradas.append(-1) | |
if salida != esperado: | |
for i in range(self.nEntradas): | |
self.pesos[i] += self.alfa*(esperado-salida)*entradas[i] | |
return (salida, (esperado-salida)**2) | |
else: | |
return (salida, (esperado-salida)**2) | |
def synapsis(self, entradas): | |
return self.funcion_no_lineal(entradas) | |
def ver_estado(self): | |
print "Neurona:", | |
for i in self.pesos: | |
print i, | |
def main(): | |
# and | |
n = Neurona(3, 0.10) | |
fl = open("entrenamiento.dat", "w") | |
for i in range(100): | |
if (i+1)%10 == 0: | |
suma = 0.0 | |
print "Ronda:", (i+1) | |
print "[Entradas]-> (Salida)" | |
temp = n.entrenar([0,0,0], 0) | |
suma += temp[1] | |
print "[0,0,0]->", temp | |
temp = n.entrenar([0,0,1], 0) | |
suma += temp[1] | |
print "[0,0,1]->", temp | |
temp = n.entrenar([0,1,0], 0) | |
suma += temp[1] | |
print "[0,1,0]->", temp | |
temp = n.entrenar([0,1,1], 0) | |
suma += temp[1] | |
print "[0,1,1]->", temp | |
temp = n.entrenar([1,0,0], 0) | |
suma += temp[1] | |
print "[1,0,0]->", temp | |
temp = n.entrenar([1,0,1], 0) | |
suma += temp[1] | |
print "[1,0,1]->", temp | |
temp = n.entrenar([1,1,0], 0) | |
suma += temp[1] | |
print "[1,1,0]->", temp | |
temp = n.entrenar([1,1,1], 1) | |
suma += temp[1] | |
print "[1,1,1]->", temp | |
print "Porcentaje de error: ", float(suma)/8 | |
fl.write(str(i)+" "+str(float(suma)/8.0)+"\n") | |
else: | |
suma = 0.0 | |
suma += n.entrenar([0,0,0], 0)[1] | |
suma += n.entrenar([0,0,1], 0)[1] | |
suma += n.entrenar([0,1,0], 0)[1] | |
suma += n.entrenar([0,1,1], 0)[1] | |
suma += n.entrenar([1,0,0], 0)[1] | |
suma += n.entrenar([1,0,1], 0)[1] | |
suma += n.entrenar([1,1,0], 0)[1] | |
suma += n.entrenar([1,1,1], 1)[1] | |
suma /= 8.0 | |
fl.write(str(i)+" "+str(suma)+"\n") | |
fl.close() | |
main() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment