stress_train.py

from neural_network import NeuralNetwork
import matplotlib.pyplot as plt
import numpy

# Setting up the value for the neural network
input_nodes = 784
hidden_nodes = 100
output_nodes = 10
learning_rate = 0.3

# Initializing the neural network
neural_network_object = NeuralNetwork(input_nodes, hidden_nodes,
                                       output_nodes, learning_rate)

scorecard = []


def scorecard_generation(record):
    test_value = record.split(',')
    network_lable = neural_network_object.query(
                                (numpy.asfarray(test_value[1:])/255.0*0.99)+0.01)
    if (int(test_value[0]) == numpy.argmax(network_lable)):
        return 1
    else:
        return 0


training_file = open('mnist_train.csv', 'r')
training_list = training_file.readlines()
training_file.close()
score = []
repetition = []

for i in range(1, 5):
    for j in range(pow(3, i)):
        for record in training_list:
            values = record.split(',')
            input_array = (numpy.asfarray(values[1:])/255.0*0.99)+0.01
            target = numpy.zeros(output_nodes)+0.01
            target[int(values[0])] = 0.99
            neural_network_object.train(input_array, target)

    test_file = open('mnist_test.csv', 'r')
    test_list = test_file.readlines()
    test_file.close()
    scorecard = map(lambda record: scorecard_generation(record), test_list)
    score.append(sum(scorecard)/10000.0)
    repetition.append(pow(3, i))


print "Accuracy:", score
print "Repetition:", repetition

# Ploting the graph Accurarcy v/s Number of runs this is an important graph as
# I found a very peculiar thing, the accuracy actually decreases with increase
# of repetition
y = score
x = repetition
plt.plot(x, y)
plt.yscale('linear')
plt.title('Multiple train run')
plt.xlabel("Nuber of runs")
plt.ylabel("Accuracy")
plt.grid(True)
plt.show()