def find_rate_learning(y_true, size_window):
good_rate = 0
min_epochs = 3000
for i in range(1000):
rate = random.random()
neuron = Window(1, y_true[:20], size_window)
error = 1
while error > (0.1**(10)):
neuron.during_epochs()
error = neuron.calculate_errors()
if neuron.epochs < min_epochs:
min_epochs = neuron.epochs
good_rate = rate
print(good_rate, min_epochs)
def main(size_window=4):
a = -1
b = 0.5
N = 20
step = (b - a) / N
size_window = size_window
input_vectors, y_true = function(a, b, N)
neuron = Window(1, y_true[:20], size_window)
error = 1
while error > (0.1**(10)):
neuron.during_epochs()
error = neuron.calculate_errors()
x_probality = [input_vectors[20]]
while x_probality[-1] < (2 * b - a):
x_probality.append(round(float(x_probality[-1] + step), 3))
print('Epochs=', neuron.epochs)
print('W=', neuron.weight)
print("Error=", neuron.error_on_last_epochs)
y_probality = neuron.neuron_prediction(21)
build_map(input_vectors, y_true, x_probality, y_probality)
