def task_Seeds():
hidden_nodes = 2
epochs = 1500
learning_rate = 0.01
momentum = 0.1
bias = True
network = NeuralNetwork(input_nodes=7,
hidden_nodes=hidden_nodes,
output_nodes=3,
learning_rate=learning_rate,
momentum=momentum,
bias=bias,
epochs=epochs)
input_list = initialize_data_with1stcolumn("seeds_dataset.csv")
output = []
indices = []
print(input_list)
for i in range(int(len(input_list) / 3)):
output.append([[1, 0, 0], input_list[i]])
indices.append(0)
for i in range(int(len(input_list) / 3)):
output.append([[0, 1, 0], input_list[i + int(len(input_list) / 3)]])
indices.append(1)
indices.append(1)
for i in range(int(len(input_list) / 3)):
output.append([[0, 0, 1],
input_list[i + int(len(input_list) / 3) * 2]])
indices.append(2)
indices.append(2)
print(output)
shuffled_output = copy.copy(output)
random.shuffle(shuffled_output)
for i in range(epochs):
for e in range(len(shuffled_output)):
network.train_manual_epochs(shuffled_output[e][1],
shuffled_output[e][0], i, e == 0)
numpy.set_printoptions(suppress=True) # avoid e-05 notation
fin = []
for i in range(len(input_list)):
fin.append(network.query(input_list[i]))
error = 0
for elem in range(len(fin)):
if numpy.argmax(fin[elem]) != indices[elem]:
error += 1
print("Seeds error rate = " + str(error / len(fin) * 100) + "%")
parameters = parameters_as_string(hidden_nodes, learning_rate, momentum,
epochs, bias)
calculate_results_table(3, indices, fin,
'Seeds result table \n' + parameters)
print_plot(network.sampling_iteration, network.errors_for_plot,
'Seeds error plot\n' + parameters)
def task_Seeds():
numpy.set_printoptions(suppress=True) # avoid e-05 notation
hidden_nodes = 4
epochs = 1000
learning_rate = 0.2
momentum = 0.9
bias = False
network = NeuralNetwork(input_nodes=7, hidden_nodes=hidden_nodes, output_nodes=3,
learning_rate=learning_rate, momentum=momentum, bias=bias, epochs=epochs)
input_list = initialize_data_with1stcolumn("data/seeds.csv")
output = []
indices = []
test = []
test = input_list[39:69] + input_list[109:139] + input_list[179:209]
for i in range(int(len(input_list) / 3)):
output.append([[1, 0, 0], input_list[i]])
for i in range(int(len(input_list) / 3)):
output.append([[0, 1, 0], input_list[i + int(len(input_list) / 3)]])
for i in range(int(len(input_list) / 3)):
output.append([[0, 0, 1], input_list[i + int(len(input_list) / 3) * 2]])
train = []
train = output[:39] + output[69:109] + output[139:179]
for i in range(0,30):
indices.append(0)
for i in range(30,60):
indices.append(1)
for i in range(60,90):
indices.append(2)
shuffled_output = copy.copy(train)
randomize = np.arange(len(train))
for i in range(len(train)):
shuffled_output[i] = shuffled_output[randomize[i]]
for i in range(epochs):
for e in range(len(shuffled_output)):
network.train_manual_epochs(shuffled_output[e][1], shuffled_output[e][0], i, e == 0)
fin = []
for i in range(len(test)):
fin.append(network.query(test[i]))
error = 0
for elem in range(len(fin)):
if numpy.argmax(fin[elem]) != indices[elem]:
error += 1
print("Seeds error rate = " + str(error / len(fin) * 100) + "%")
parameters = parameters_as_string(hidden_nodes, learning_rate, momentum, epochs, bias)
calculate_results_table(3, indices, fin, 'Seeds result table \n' + parameters)
print_plot(network.sampling_iteration, network.errors_for_plot, 'Zbiór: nasiona\n Wykres zależności wartości błędu od ilości epok\n' + parameters)
def task_SISE():
epochs = 200
hidden_nodes = 5
learning_rate = 0.2
momentum = 0.9
bias = True
network = NeuralNetwork(input_nodes=1,
hidden_nodes=hidden_nodes,
output_nodes=1,
learning_rate=learning_rate,
momentum=momentum,
bias=bias,
epochs=epochs)
input_list = list(range(1, 100))
output = []
for var in input_list:
output.append(math.sqrt(var))
c = list(zip(input_list, output))
# shuffle in the same way
random.shuffle(c)
input_list, output = zip(*c)
input_norm = normalize(input_list)
output_norm = normalize(output)
for e in range(1000):
for i in range(len(input_norm)):
network.train_manual_epochs([input_norm[i]], [output_norm[i]],
e) # inputs need to be arrays
fin = []
query_list = []
for i in range(len(input_norm)):
query_list.append(input_norm[i])
fin.append(network.query([query_list[i]]))
numpy.set_printoptions(suppress=True) # avoid e-05 notation
print(output)
results = denormalize(output, fin)
for var in results:
print(var)
result_tab = numpy.zeros(shape=(len(input_norm), len(input_norm)))
for i in range(len(fin)):
result_tab[numpy.argmax(query_list[i])][numpy.argmax(fin[i])] += 1
print('Ex 1 results table')
print(result_tab)
parameters = parameters_as_string(hidden_nodes, learning_rate, momentum,
epochs, bias)
funs.print_plot(network.sampling_iteration, network.errors_for_plot,
'Ex1 error plot ' + parameters)
def task_2():
epochs = 200
hidden_nodes = 10
learning_rate = 0.2
momentum = 0.9
bias = True
network = NeuralNetwork(input_nodes=4,
hidden_nodes=hidden_nodes,
output_nodes=4,
learning_rate=learning_rate,
momentum=momentum,
bias=bias,
epochs=epochs)
input_list = [1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], [0, 0, 0, 1]
output = [1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 1, 0], [0, 0, 0, 1]
x = numpy.zeros(epochs * 4, dtype=int)
for i in range(epochs):
x[i] = 0
x[i + epochs] = 1
x[i + epochs * 2] = 2
x[i + epochs * 3] = 3
random.shuffle(x)
for e in range(len(x)):
network.train_manual_epochs(input_list[(x[e])], output[(x[e])], e)
fin = []
query_list = []
for i in range(64):
query_list.append(input_list[x[i]])
fin.append(network.query(query_list[i]))
numpy.set_printoptions(suppress=True) # avoid e-05 notation
print(output)
for elem in range(len(fin)):
print(fin[elem])
result_tab = numpy.zeros(shape=(len(input_list), len(input_list)))
for i in range(len(fin)):
result_tab[numpy.argmax(query_list[i])][numpy.argmax(fin[i])] += 1
print('Ex 1 results table')
print(result_tab)
parameters = parameters_as_string(hidden_nodes, learning_rate, momentum,
epochs, bias)
funs.print_plot(network.sampling_iteration, network.errors_for_plot,
'Ex1 error plot ' + parameters)
def task_SISE():
epochs = 2000
hidden_nodes = 5
learning_rate = 0.2
momentum = 0.9
bias = False
network = NeuralNetwork(input_nodes=1,
hidden_nodes=hidden_nodes,
output_nodes=1,
learning_rate=learning_rate,
momentum=momentum,
bias=bias,
epochs=epochs)
input_list = []
input_list = list(range(1, 100))
output = []
for var in input_list:
output.append(math.sqrt(var))
c = list(zip(input_list, output))
# shuffle in the same way
random.shuffle(c)
input_list, output = zip(*c)
input_norm = normalize(input_list)
output_norm = normalize(output)
for e in range(epochs):
for i in range(len(input_norm)):
network.train_manual_epochs([input_norm[i]], [output_norm[i]],
e) # inputs need to be arrays
fin = []
query_list = []
for i in range(len(input_norm)):
query_list.append(input_norm[i])
fin.append(network.query([query_list[i]]))
numpy.set_printoptions(suppress=True) # avoid e-05 notation
print(output)
results = denormalize(output, fin)
for e in range(len(results)):
print(input_list[e], results[e])
result_tab = numpy.zeros(shape=(len(input_norm), len(input_norm)))
for i in range(len(fin)):
result_tab[numpy.argmax(query_list[i])][numpy.argmax(fin[i])] += 1
print('Results table')
print(result_tab)
parameters = parameters_as_string(hidden_nodes, learning_rate, momentum,
epochs, bias)
funs.print_plot(network.sampling_iteration, network.errors_for_plot,
'Error plot ' + parameters)
flat_results = [item for sublist in results for item in sublist]
flat_input = list(input_list)
flat_output = list(output)
title = "Expected(green) and calculated(red) values"
pyplot.figure(title)
pyplot.plot(flat_input, flat_results, 'r.')
pyplot.plot(flat_input, flat_output, 'g.')
pyplot.grid(axis='both',
color='black',
which='major',
linestyle='--',
linewidth=1)
pyplot.xlabel("x")
pyplot.ylabel("sqrt(x)")
pyplot.suptitle(title)
pyplot.savefig('plotValues.png')
pyplot.show()
def task_MNIST():
numpy.set_printoptions(suppress=True) # avoid e-05 notation
hidden_nodes = 10
epochs = 3
learning_rate = 0.05
momentum = 0.2
bias = True
network = NeuralNetwork(input_nodes=784,
hidden_nodes=hidden_nodes,
output_nodes=10,
learning_rate=learning_rate,
momentum=momentum,
bias=bias,
epochs=epochs,
error_sampling_rate=(1 / epochs))
# load the mnist training data CSV file into a list
training_data_file = open("mnist_train.csv", 'r')
training_data_list = training_data_file.readlines()
training_data_file.close()
# train the neural network
# go through all records in the training data set
for e in range(epochs):
print("Epoch:" + str(e))
collect_data_for_plot = True
for record in training_data_list:
# split the record by the ',' commas
all_values = record.split(',')
# scale and shift the inputs
inputs = (numpy.asfarray(all_values[1:]) / 255.0 * 0.99) + 0.01
# create the target output values (all 0.01, except the desired label which is 0.99)
targets = numpy.zeros(10) + 0.01
# all_values[0] is the target label for this record
targets[int(all_values[0])] = 0.99
network.train_manual_epochs(inputs.tolist(), targets, e,
collect_data_for_plot)
collect_data_for_plot = False
# load the mnist test data CSV file into a list
test_data_file = open("mnist_test.csv", 'r')
test_data_list = test_data_file.readlines()
test_data_file.close()
error = 0
# test the neural network
# scorecard for how well the network performs, initially empty
scorecard = []
# go through all the records in the test data set
fin = []
values = []
for record in test_data_list:
# split the record by the ',' commas
all_values = record.split(',')
# correct answer is first value
correct_label = int(all_values[0])
values.append(correct_label)
# print(correct_label, "correct label")
# scale and shift the inputs
inputs = (numpy.asfarray(all_values[1:]) / 255.0 * 0.99) + 0.01
# query the network
outputs = network.query(inputs.tolist())
fin.append(outputs)
# the index of the highest value corresponds to the label
label = numpy.argmax(outputs)
# print(label, "network's answer")
# append correct or incorrect to list
if label == correct_label:
# network's answer matches correct answer, add 1 to scorecard
error += 1
print("MNIST accuracy = " + str(error / 10000 * 100) + "%")
parameters = parameters_as_string(hidden_nodes, learning_rate, momentum,
epochs, bias)
calculate_results_table(10, values, fin,
'MNIST result table\n' + parameters)
print_plot(network.sampling_iteration, network.errors_for_plot,
'MNIST error plot \n' + parameters)