with open('seeddata.csv', newline='') as csvfile:
rows = reader(csvfile, delimiter=' ')
Data = [row[0].split(',') for row in rows]
NN = NeuralNetwork()
Network = NN.Initialize(Data)
for i in range(1, 10000):
shuffle(Data)
sum_error = 0
for data in Data:
NN.InitializeInputs(data, Network)
NN.FowardPropagate(Network)
expected = NN.Expected(data[-1])
####################################################################
outputs = [node.Activation for node in Network[-1].Neurons]
outputLayer = Network[-1].Neurons
maxindex = reduce(
lambda x, y: x
if outputLayer[x].Activation > outputLayer[y].Activation else y,
range(len(outputLayer)))
if maxindex == NN.OutcomeToIndex[data[-1]]:
sum_error += 1
# sum_error += sum([(expected[i]-outputs[i])**2 for i in range(len(expected))])
#############################################################################
NN.BackwardPropagate(expected, Network)
NN.Train(Network)
print(sum_error / len(Data) * 100)
