gistfile1.py

import math

class Neuron:

	def __init__(self, network, ID, aE, threshhold = 0, isBias = 0):
		self.network = network
		self.weights = {}
		self.successors = []
		self.values = []
		self.ID = ID
		self.activationEquation = aE
		if not isBias:
			network.addSynapse(0, ID, -threshhold)

	def __str__(self):
		return 'Neuron(' + self.ID + ', Successors(' + str(self.successors) + ')'

	def feed(self, successor):
		self.successors.append(successor)

	def poll(self, ancestor, weight):
		self.weights[ancestor] = weight

	def update(self, ID, net):
		self.values.append(weights[ID] * net)
		if len(self.weights) == len(self.values):
			propagate()

	def propagate(self):
		net = 0
		for value in self.values:
			net += value
		activate(net)

	def activate(self, net):
		push(self.activationEquation(net))

	def push(self, net):
		for successor in self.successors:
			self.net.neurons[successor].update(ID, net)
		values = []

class ActivationEquations:

	def fermi(self, net):
		return 1 / (1 + math.exp(-net * 25))

	def binary(self, net):
		if net > 0:
			return 1
		return 0

	def one(self, net):
		return 1