Project

Predicting Hand Written Numbers using Neural Network

KKeerthidhar Reddy3 Aug 2022

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import keras

In [118]:

df=pd.read_csv('mnist_train.csv')
df1=pd.read_csv('mnist_test.csv')

In [119]:

df

Out[119]:

	label	1x1	1x2	1x3	1x4	1x5	1x6	1x7	1x8	1x9	...	28x19	28x20	28x21	28x22	28x23	28x24	28x25	28x26	28x27	28x28
0	5	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
2	4	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
3	1	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
4	9	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
59995	8	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
59996	3	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
59997	5	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
59998	6	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
59999	8	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0

60000 rows × 785 columns

In [120]:

df1

Out[120]:

	label	1x1	1x2	1x3	1x4	1x5	1x6	1x7	1x8	1x9	...	28x19	28x20	28x21	28x22	28x23	28x24	28x25	28x26	28x27	28x28
0	7	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
1	2	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
2	1	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
4	4	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
9995	2	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
9996	3	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
9997	4	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
9998	5	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0
9999	6	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0

10000 rows × 785 columns

In [122]:

X_train=df.drop(['label'],axis=1)
y_train=df['label']

In [124]:

X_test=df1.drop(['label'],axis=1)
y_test=df1['label']

In [125]:

X_train.shape,y_train.shape,X_test.shape,y_test.shape

Out[125]:

((60000, 784), (60000,), (10000, 784), (10000,))

In [126]:

#Expand the dimension of the image

In [127]:

X_train = X_train.values.astype('float32')
X_train = X_train.reshape(X_train.shape[0],28, 28,1)

In [128]:

X_train.shape

Out[128]:

(60000, 28, 28, 1)

In [129]:

X_test = X_test.values.astype('float32')
X_test = X_test.reshape(X_test.shape[0],28, 28,1)

In [130]:

X_test.shape

Out[130]:

(10000, 28, 28, 1)

In [ ]:

# image plotting

In [121]:

def plot_image(n):
    plt.imshow(X_train[n])
    print('The Number is:',y_train[n])
    plt.show()

In [101]:

#Normalizing the image

In [102]:

X_train=X_train/255
X_test=X_test/255

In [131]:

for i in range(0,5):
    plot_image(i)

The Number is: 5

The Number is: 0

The Number is: 4

The Number is: 1

The Number is: 9

In [ ]:

#Converting to one hot vectors

In [132]:

y_train=keras.utils.to_categorical(y_train)

In [134]:

y_test=keras.utils.to_categorical(y_test)

In [136]:

y_train

Out[136]:

array([[0., 0., 0., ..., 0., 0., 0.],
       [1., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       ...,
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 1., 0.]], dtype=float32)

In [137]:

y_test

Out[137]:

array([[0., 0., 0., ..., 1., 0., 0.],
       [0., 0., 1., ..., 0., 0., 0.],
       [0., 1., 0., ..., 0., 0., 0.],
       ...,
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.]], dtype=float32)

In [138]:

from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Conv2D, MaxPool2D

In [142]:

model=Sequential()
model.add(Conv2D(32,(3,3),input_shape=(28,28,1),activation='relu'))
model.add(MaxPool2D(2,2))

model.add(Conv2D(64,(3,3),activation='relu'))
model.add(MaxPool2D(2,2))

model.add(Flatten())

model.add(Dropout(0.25))

model.add(Dense(10,activation='softmax'))

In [143]:

model.summary()

Model: "sequential_3"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 conv2d_4 (Conv2D)           (None, 26, 26, 32)        320       
                                                                 
 max_pooling2d_2 (MaxPooling  (None, 13, 13, 32)       0         
 2D)                                                             
                                                                 
 conv2d_5 (Conv2D)           (None, 11, 11, 64)        18496     
                                                                 
 max_pooling2d_3 (MaxPooling  (None, 5, 5, 64)         0         
 2D)                                                             
                                                                 
 flatten_1 (Flatten)         (None, 1600)              0         
                                                                 
 dropout (Dropout)           (None, 1600)              0         
                                                                 
 dense (Dense)               (None, 10)                16010     
                                                                 
=================================================================
Total params: 34,826
Trainable params: 34,826
Non-trainable params: 0
_________________________________________________________________

In [145]:

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

In [146]:

img=model.fit(X_train, y_train,epochs=20, verbose=1)

Epoch 1/20
1875/1875 [==============================] - 33s 15ms/step - loss: 0.4093 - accuracy: 0.9299
Epoch 2/20
1875/1875 [==============================] - 28s 15ms/step - loss: 0.0935 - accuracy: 0.9727
Epoch 3/20
1875/1875 [==============================] - 27s 14ms/step - loss: 0.0761 - accuracy: 0.9770
Epoch 4/20
1875/1875 [==============================] - 27s 14ms/step - loss: 0.0667 - accuracy: 0.9800
Epoch 5/20
1875/1875 [==============================] - 27s 14ms/step - loss: 0.0620 - accuracy: 0.9815
Epoch 6/20
1875/1875 [==============================] - 25s 13ms/step - loss: 0.0543 - accuracy: 0.9840
Epoch 7/20
1875/1875 [==============================] - 25s 13ms/step - loss: 0.0543 - accuracy: 0.9839
Epoch 8/20
1875/1875 [==============================] - 25s 13ms/step - loss: 0.0493 - accuracy: 0.9855
Epoch 9/20
1875/1875 [==============================] - 26s 14ms/step - loss: 0.0470 - accuracy: 0.9860
Epoch 10/20
1875/1875 [==============================] - 626s 334ms/step - loss: 0.0442 - accuracy: 0.9878
Epoch 11/20
1875/1875 [==============================] - 25s 13ms/step - loss: 0.0425 - accuracy: 0.9876
Epoch 12/20
1875/1875 [==============================] - 25s 13ms/step - loss: 0.0419 - accuracy: 0.9881
Epoch 13/20
1875/1875 [==============================] - 24s 13ms/step - loss: 0.0407 - accuracy: 0.9882
Epoch 14/20
1875/1875 [==============================] - 25s 13ms/step - loss: 0.0412 - accuracy: 0.9889
Epoch 15/20
1875/1875 [==============================] - 28s 15ms/step - loss: 0.0395 - accuracy: 0.9893
Epoch 16/20
1875/1875 [==============================] - 28s 15ms/step - loss: 0.0388 - accuracy: 0.9886
Epoch 17/20
1875/1875 [==============================] - 28s 15ms/step - loss: 0.0391 - accuracy: 0.9896
Epoch 18/20
1875/1875 [==============================] - 26s 14ms/step - loss: 0.0371 - accuracy: 0.9900
Epoch 19/20
1875/1875 [==============================] - 29s 15ms/step - loss: 0.0358 - accuracy: 0.9899
Epoch 20/20
1875/1875 [==============================] - 26s 14ms/step - loss: 0.0403 - accuracy: 0.9902

In [148]:

predictions = model.predict(X_test, verbose=0)

In [150]:

score=model.evaluate(X_test,y_test)

313/313 [==============================] - 2s 5ms/step - loss: 0.0496 - accuracy: 0.9881

In [151]:

print('The model accuracy is:',score[1])

The model accuracy is: 0.988099992275238

In [ ]:

Predicting Hand Written Numbers using Neural Network

Ship work like this
in your own course.

Talk to a career counsellor

Ship work like thisin your own course.

Talk to a career counsellor

Ship work like this
in your own course.