9/30/23, 11:13 AM Untitled0.

ipynb - Colaboratory

from tensorflow.keras.datasets import mnist

# Load MNIST dataset

(x_train, y_train), (x_test, y_test) = mnist.load_data()

# Print the shape of the loaded data to verify it's loaded correctly
print("Training data shape:", x_train.shape)
print("Training labels shape:", y_train.shape)
print("Test data shape:", x_test.shape)
print("Test labels shape:", y_test.shape)

Training data shape: (60000, 28, 28)

Training labels shape: (60000,)
Test data shape: (10000, 28, 28)
Test labels shape: (10000,)

from tensorflow.keras.datasets import mnist

import matplotlib.pyplot as plt

# Load MNIST dataset

(x_train, y_train), (x_test, y_test) = mnist.load_data()

# Display a few sample images

plt.figure(figsize=(12, 4))
for i in range(10):
plt.subplot(2, 5, i+1)
plt.imshow(x_train[i], cmap='gray')
plt.title(f"Label: {y_train[i]}")
plt.axis('off')
plt.show()

output

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense

# Load MNIST dataset

(x_train, y_train), (x_test, y_test) = mnist.load_data()

# Normalize the pixel values to be between 0 and 1

x_train = x_train.reshape((-1, 28, 28, 1)).astype('float32') / 255.0
x_test = x_test.reshape((-1, 28, 28, 1)).astype('float32') / 255.0

# Convert labels to one-hot encoding

y_train = tf.keras.utils.to_categorical(y_train, 10)
y_test = tf.keras.utils.to_categorical(y_test, 10)

# Create a custom CNN model

model = Sequential([
Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)),
MaxPooling2D((2, 2)),
Conv2D(64, (3, 3), activation='relu'),
MaxPooling2D((2, 2)),
Flatten(),
Dense(128, activation='relu'),
Dense(10, activation='softmax')
])

# Compile the model

https://colab.research.google.com/drive/1uOafKc41edpKxQwJ0yxQxNUXZlNDJUD5#scrollTo=oGBZADsWh3_i&printMode=true 1/5
9/30/23, 11:13 AM Untitled0.ipynb - Colaboratory
model.compile(optimizer='adam',
loss='categorical_crossentropy',
metrics=['accuracy'])

# Print the summary of the model

model.summary()

# Train the model

model.fit(x_train, y_train, epochs=5, batch_size=64, validation_split=0.2)

# Evaluate the model on test data

test_loss, test_accuracy = model.evaluate(x_test, y_test)
print(f'Test accuracy: {test_accuracy}')

Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d (Conv2D) (None, 26, 26, 32) 320

max_pooling2d (MaxPooling2 (None, 13, 13, 32) 0

D)

conv2d_1 (Conv2D) (None, 11, 11, 64) 18496

max_pooling2d_1 (MaxPoolin (None, 5, 5, 64) 0

g2D)

flatten (Flatten) (None, 1600) 0

dense (Dense) (None, 128) 204928

dense_1 (Dense) (None, 10) 1290

=================================================================
Total params: 225034 (879.04 KB)
Trainable params: 225034 (879.04 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________
Epoch 1/5
750/750 [==============================] - 34s 43ms/step - loss: 0.1832 - accuracy: 0.9458 - val_loss: 0.0673 - val_accuracy: 0.979
Epoch 2/5
750/750 [==============================] - 27s 36ms/step - loss: 0.0555 - accuracy: 0.9822 - val_loss: 0.0469 - val_accuracy: 0.985
Epoch 3/5
750/750 [==============================] - 26s 35ms/step - loss: 0.0356 - accuracy: 0.9889 - val_loss: 0.0470 - val_accuracy: 0.986
Epoch 4/5
750/750 [==============================] - 27s 36ms/step - loss: 0.0274 - accuracy: 0.9914 - val_loss: 0.0439 - val_accuracy: 0.987
Epoch 5/5
750/750 [==============================] - 26s 35ms/step - loss: 0.0211 - accuracy: 0.9930 - val_loss: 0.0396 - val_accuracy: 0.989
313/313 [==============================] - 2s 6ms/step - loss: 0.0279 - accuracy: 0.9900
Test accuracy: 0.9900000095367432

import numpy as np
from tensorflow.keras.datasets import mnist
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense
from tensorflow.keras.utils import to_categorical

# Load MNIST dataset

(x_train, y_train), (x_test, y_test) = mnist.load_data()

# Preprocess the data

x_train = x_train.reshape(x_train.shape[0], 28, 28, 1).astype('float32') / 255
x_test = x_test.reshape(x_test.shape[0], 28, 28, 1).astype('float32') / 255
y_train = to_categorical(y_train, 10)
y_test = to_categorical(y_test, 10)

# Define the CNN model

model = Sequential([
Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 1)),
MaxPooling2D(pool_size=(2, 2)),
Conv2D(64, kernel_size=(3, 3), activation='relu'),
MaxPooling2D(pool_size=(2, 2)),
Flatten(),
Dense(128, activation='relu'),
Dense(10, activation='softmax')
])

# Compile the model

model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

# Train the CNN model

model.fit(x_train, y_train, epochs=5, batch_size=128, validation_split=0.2)

https://colab.research.google.com/drive/1uOafKc41edpKxQwJ0yxQxNUXZlNDJUD5#scrollTo=oGBZADsWh3_i&printMode=true 2/5
9/30/23, 11:13 AM Untitled0.ipynb - Colaboratory

# Define a new model with the same input as the original model
feature_extraction_model = Model(inputs=model.input, outputs=model.layers[4].output) # Choose the appropriate layer for feature extractio

# Extract features from the CNN model

cnn_features_train = feature_extraction_model.predict(x_train)
cnn_features_test = feature_extraction_model.predict(x_test)

# Print the shape of the extracted features

print("Shape of CNN features (train):", cnn_features_train.shape)
print("Shape of CNN features (test):", cnn_features_test.shape)

Epoch 1/5
375/375 [==============================] - 28s 72ms/step - loss: 0.2438 - accuracy: 0.9306 - val_loss: 0.0879 - val_accuracy: 0.972
Epoch 2/5
375/375 [==============================] - 24s 63ms/step - loss: 0.0643 - accuracy: 0.9805 - val_loss: 0.0529 - val_accuracy: 0.983
Epoch 3/5
375/375 [==============================] - 25s 67ms/step - loss: 0.0435 - accuracy: 0.9866 - val_loss: 0.0469 - val_accuracy: 0.986
Epoch 4/5
375/375 [==============================] - 26s 68ms/step - loss: 0.0346 - accuracy: 0.9892 - val_loss: 0.0431 - val_accuracy: 0.986
Epoch 5/5
375/375 [==============================] - 25s 67ms/step - loss: 0.0263 - accuracy: 0.9919 - val_loss: 0.0505 - val_accuracy: 0.985
1875/1875 [==============================] - 11s 6ms/step
313/313 [==============================] - 2s 6ms/step
Shape of CNN features (train): (60000, 1600)
Shape of CNN features (test): (10000, 1600)

from sklearn.svm import SVC

from sklearn.metrics import accuracy_score, confusion_matrix

# Assuming cnn_features_train and cnn_features_test are extracted features from the CNN model

# Initialize and train the SVM classifier

svm_classifier = SVC(kernel='linear')
svm_classifier.fit(cnn_features_train, np.argmax(y_train, axis=1))

# Predict using the trained SVM classifier

svm_predictions = svm_classifier.predict(cnn_features_test)

# Calculate accuracy
svm_accuracy = accuracy_score(np.argmax(y_test, axis=1), svm_predictions)
print("SVM Classifier Accuracy:", svm_accuracy)

# Confusion matrix
svm_conf_matrix = confusion_matrix(np.argmax(y_test, axis=1), svm_predictions)
print("Confusion Matrix for SVM Classifier:")
print(svm_conf_matrix)

SVM Classifier Accuracy: 0.9915

Confusion Matrix for SVM Classifier:
[[ 978 0 0 0 0 0 0 1 1 0]
[ 0 1134 0 0 0 0 0 1 0 0]
[ 1 1 1020 1 1 0 0 5 2 1]
[ 0 0 3 1002 0 3 0 1 1 0]
[ 0 0 0 0 977 0 1 0 0 4]
[ 1 0 1 6 0 880 2 1 1 0]
[ 3 3 1 0 1 4 944 0 2 0]
[ 0 4 1 0 0 0 0 1023 0 0]
[ 3 0 2 0 0 0 0 1 965 3]
[ 0 1 0 1 8 5 0 2 0 992]]

# Evaluate the trained SVM classifier

svm_accuracy = accuracy_score(np.argmax(y_test, axis=1), svm_predictions)
svm_conf_matrix = confusion_matrix(np.argmax(y_test, axis=1), svm_predictions)
print("SVM Classifier Accuracy:", svm_accuracy)
print("Confusion Matrix for SVM Classifier:")
print(svm_conf_matrix)

# Evaluate the trained CNN model

cnn_test_loss, cnn_test_accuracy = model.evaluate(x_test, y_test)
print("CNN Model Accuracy on Test Data:", cnn_test_accuracy)

SVM Classifier Accuracy: 0.9915

Confusion Matrix for SVM Classifier:
[[ 978 0 0 0 0 0 0 1 1 0]
[ 0 1134 0 0 0 0 0 1 0 0]
[ 1 1 1020 1 1 0 0 5 2 1]
[ 0 0 3 1002 0 3 0 1 1 0]
[ 0 0 0 0 977 0 1 0 0 4]
[ 1 0 1 6 0 880 2 1 1 0]
[ 3 3 1 0 1 4 944 0 2 0]

https://colab.research.google.com/drive/1uOafKc41edpKxQwJ0yxQxNUXZlNDJUD5#scrollTo=oGBZADsWh3_i&printMode=true 3/5
9/30/23, 11:13 AM Untitled0.ipynb - Colaboratory
[ 0 4 1 0 0 0 0 1023 0 0]
[ 3 0 2 0 0 0 0 1 965 3]
[ 0 1 0 1 8 5 0 2 0 992]]
313/313 [==============================] - 4s 13ms/step - loss: 0.0414 - accuracy: 0.9859
CNN Model Accuracy on Test Data: 0.9858999848365784

import numpy as np
from sklearn.metrics import confusion_matrix

# Assuming model is the trained CNN model and x_test, y_test are the test data
cnn_predictions = model.predict(x_test) # Get raw predictions
cnn_classes = np.argmax(cnn_predictions, axis=1) # Get predicted classes

# Calculate and print the confusion matrix for CNN predictions

cnn_conf_matrix = confusion_matrix(np.argmax(y_test, axis=1), cnn_classes)
print("Confusion Matrix for CNN Model:")
print(cnn_conf_matrix)

313/313 [==============================] - 2s 8ms/step

Confusion Matrix for CNN Model:
[[ 974 0 0 0 0 0 2 1 3 0]
[ 0 1117 4 0 0 0 5 0 9 0]
[ 2 0 1027 0 0 0 0 1 2 0]
[ 0 0 2 1000 0 5 0 1 2 0]
[ 0 0 0 0 963 0 1 0 1 17]
[ 2 0 1 9 0 868 3 0 5 4]
[ 1 2 1 0 1 1 947 0 5 0]
[ 1 3 11 1 0 0 0 994 2 16]
[ 2 0 2 1 0 0 0 0 966 3]
[ 1 0 0 0 1 0 0 0 4 1003]]

import seaborn as sns

import matplotlib.pyplot as plt

# Assuming svm_conf_matrix and cnn_conf_matrix are the confusion matrices calculated previously

# Generate heatmap for SVM Classifier

plt.figure(figsize=(8, 6))
sns.heatmap(svm_conf_matrix, annot=True, fmt='d', cmap='Blues')
plt.xlabel('Predicted Labels')
plt.ylabel('True Labels')
plt.title('Confusion Matrix for SVM Classifier')
plt.show()

# Generate heatmap for CNN Model

plt.figure(figsize=(8, 6))
sns.heatmap(cnn_conf_matrix, annot=True, fmt='d', cmap='Blues')
plt.xlabel('Predicted Labels')
plt.ylabel('True Labels')
plt.title('Confusion Matrix for CNN Model')
plt.show()

https://colab.research.google.com/drive/1uOafKc41edpKxQwJ0yxQxNUXZlNDJUD5#scrollTo=oGBZADsWh3_i&printMode=true 4/5
9/30/23, 11:13 AM Untitled0.ipynb - Colaboratory

https://colab.research.google.com/drive/1uOafKc41edpKxQwJ0yxQxNUXZlNDJUD5#scrollTo=oGBZADsWh3_i&printMode=true 5/5