In [3]:

import numpy as np 
import pandas as pd
import matplotlib.pyplot as plt
import tensorflow as tf

from sklearn.model_selection import train_test_split
from tensorflow.python import keras
from tensorflow.python.keras.models import Sequential
from tensorflow.python.keras.layers import Dense, Flatten, Conv2D, Dropout

In [22]:

# Import Data
train = pd.read_csv("train.csv")
test = pd.read_csv("test.csv")

num_classes = 10
img_rows = img_cols = 28

def data_prep(raw):
    out_y = keras.utils.to_categorical(raw.label, num_classes)
    num_images = raw.shape[0]
    x_as_array = raw.values[:,1:]
    x_shaped_array = x_as_array.reshape(num_images, img_rows, img_cols, 1)
    out_x = x_shaped_array / 255
    return out_x, out_y

x_train, y_train = data_prep(train)

x_test = test.values.reshape(test.shape[0], img_rows, img_cols, 1)

In [23]:

model = Sequential()
model.add(Conv2D(20, kernel_size=(3, 3),
                 activation='relu',
                 input_shape=(img_rows, img_cols, 1)))
model.add(Conv2D(20, kernel_size=(3, 3), activation='relu'))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dense(num_classes, activation='softmax'))

model.compile(loss=keras.losses.categorical_crossentropy,
              optimizer='adam',
              metrics=['accuracy'])
model.fit(x_train, y_train,
          batch_size=128,
          epochs=2,
          validation_split = 0.2)

Train on 33600 samples, validate on 8400 samples
Epoch 1/2
33600/33600 [==============================] - 82s 2ms/sample - loss: 0.2185 - acc: 0.9376 - val_loss: 0.0815 - val_acc: 0.9730
Epoch 2/2
33600/33600 [==============================] - 82s 2ms/sample - loss: 0.0600 - acc: 0.9822 - val_loss: 0.0567 - val_acc: 0.9829

Out[23]:

<tensorflow.python.keras.callbacks.History at 0x24c1071c438>

In [97]:

preds = pd.DataFrame(model.predict(x_test))

def decode(row):
    dev = []
    for c in preds.columns:
        dev.append(abs(1 - row[c]))
    return(np.where(dev == np.amin(dev))[0][0])
        
output = pd.DataFrame(preds.apply(decode, axis = 1), dtype = "int", columns = ["Label"])
output.index.name = "ImageId"
output.index += 1

output.to_csv('submission.csv')

In [95]:

output
# arr = [2, 3, 4, 1]
# np.where(arr == np.amin(arr))[0][0]

Out[95]:

	Label
ImageId
1	2
2	0
3	9
4	0
5	3
6	7
7	0
8	3
9	0
10	3
11	5
12	7
13	4
14	0
15	4
16	3
17	3
18	1
19	9
20	0
21	9
22	1
23	1
24	5
25	7
26	4
27	2
28	7
29	4
30	7
...	...
27971	5
27972	0
27973	4
27974	8
27975	0
27976	3
27977	6
27978	0
27979	1
27980	9
27981	3
27982	1
27983	1
27984	0
27985	4
27986	5
27987	2
27988	2
27989	9
27990	6
27991	7
27992	6
27993	1
27994	9
27995	7
27996	9
27997	7
27998	3
27999	9
28000	2

28000 rows × 1 columns