模型-MLP分類法One Hot Encoding

t=2
category=2
dim=x_train.shape[1]


y_train2=tf.keras.utils.to_categorical(y_train, num_classes=(category))
y_test2=tf.keras.utils.to_categorical(y_test, num_classes=(category))
# 建立模型
model = tf.keras.models.Sequential()
model.add(tf.keras.layers.Dense(units=200,
    activation=tf.nn.relu,
    input_dim=dim))
model.add(tf.keras.layers.Dense(units=40*t,
    activation=tf.nn.relu ))
model.add(tf.keras.layers.Dense(units=80*t,
    activation=tf.nn.relu ))
model.add(tf.keras.layers.Dense(units=100*t,
    activation=tf.nn.relu ))
model.add(tf.keras.layers.Dense(units=category,
    activation=tf.nn.softmax ))
model.compile(optimizer='adam',
    loss=tf.keras.losses.categorical_crossentropy,
    metrics=['accuracy'])
model.fit(x_train, y_train2,
          epochs=300*t,
          batch_size=64)

#測試
model.summary()

score = model.evaluate(x_test, y_test2, batch_size=64)
print("score:",score)

predict2 = model.predict_classes(x_test)
print("predict_classes:",predict2)
print("y_test",y_test[:])
Python
