[머신러닝스터디/2016]
[머신러닝스터디/2016/목차]
== 내용 ==
 * Project
  * [https://www.kaggle.com/c/digit-recognizer kaggle digit recignizer]
 * 지난 시간( [머신러닝스터디/2016/2016_09_03]) 학습이 안되는 문제 해결
  * Normalization 이용
    * `X_train/255.` 부분이 Normalization
    * feature의 범위를 0~255에서 0~1 사이의 실수로 scaling 함. 
=== 코드 ===
{{{
import pandas as pd
import keras
from sklearn.cross_validation import train_test_split
from keras.utils.np_utils import to_categorical

train = pd.read_csv("../input/train.csv")
test  = pd.read_csv("../input/test.csv")

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

X_train, X_test, y_train, y_test = train_test_split(X_train/255., y_train, test_size=0.30)

model = keras.models.Sequential()

model.add(keras.layers.Dense(64, input_dim=28*28, activation='relu'))
model.add(keras.layers.Dense(32, activation='relu'))
model.add(keras.layers.Dense(16, activation='relu'))
model.add(keras.layers.Dense(10, activation='softmax'))

model.compile(loss='categorical_crossentropy', optimizer='adagrad', metrics=['accuracy'])
model.fit(X_train, to_categorical(y_train, 10), nb_epoch=5, batch_size=600)

score = model.evaluate(X_test, to_categorical(y_test, 10), batch_size=700)

print(score)

print(model.predict(X_test))[0]
print(y_test[0])
}}}
== 다음 시간에는 ==
 * 한 주 쉬어갑니다.