- Yapay sinir ağları ile ilgili slaytı indirmek için tıklayınız.
Yakıt Verimliliği Çalışması
Bu çalışmada yakıt verimliliği ile aşağıdaki sayfada verilen eğitim üzerine çalışmalar yapılmıştır.
https://www.tensorflow.org/tutorials/keras/regression
Veri setini indirmek için tıklayınız.
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd # pip install pandas
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
column_names = ['MPG', 'Cylinders', 'Displacement', 'Horsepower', 'Weight',
'Acceleration', 'Model Year', 'Origin']
dataset=pd.read_csv("04_auto-mpg.data", names=column_names, na_values='?',
comment='\t', sep=' ', skipinitialspace=True)
dataset
| MPG | Cylinders | Displacement | Horsepower | Weight | Acceleration | Model Year | Origin | |
|---|---|---|---|---|---|---|---|---|
| 0 | 18.0 | 8 | 307.0 | 130.0 | 3504.0 | 12.0 | 70 | 1 |
| 1 | 15.0 | 8 | 350.0 | 165.0 | 3693.0 | 11.5 | 70 | 1 |
| 2 | 18.0 | 8 | 318.0 | 150.0 | 3436.0 | 11.0 | 70 | 1 |
| 3 | 16.0 | 8 | 304.0 | 150.0 | 3433.0 | 12.0 | 70 | 1 |
| 4 | 17.0 | 8 | 302.0 | 140.0 | 3449.0 | 10.5 | 70 | 1 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 393 | 27.0 | 4 | 140.0 | 86.0 | 2790.0 | 15.6 | 82 | 1 |
| 394 | 44.0 | 4 | 97.0 | 52.0 | 2130.0 | 24.6 | 82 | 2 |
| 395 | 32.0 | 4 | 135.0 | 84.0 | 2295.0 | 11.6 | 82 | 1 |
| 396 | 28.0 | 4 | 120.0 | 79.0 | 2625.0 | 18.6 | 82 | 1 |
| 397 | 31.0 | 4 | 119.0 | 82.0 | 2720.0 | 19.4 | 82 | 1 |
398 rows × 8 columns
dataset.isna().sum()
MPG 0
Cylinders 0
Displacement 0
Horsepower 6
Weight 0
Acceleration 0
Model Year 0
Origin 0
dtype: int64
# clean data
dataset = dataset.dropna()
dataset
| MPG | Cylinders | Displacement | Horsepower | Weight | Acceleration | Model Year | Origin | |
|---|---|---|---|---|---|---|---|---|
| 0 | 18.0 | 8 | 307.0 | 130.0 | 3504.0 | 12.0 | 70 | 1 |
| 1 | 15.0 | 8 | 350.0 | 165.0 | 3693.0 | 11.5 | 70 | 1 |
| 2 | 18.0 | 8 | 318.0 | 150.0 | 3436.0 | 11.0 | 70 | 1 |
| 3 | 16.0 | 8 | 304.0 | 150.0 | 3433.0 | 12.0 | 70 | 1 |
| 4 | 17.0 | 8 | 302.0 | 140.0 | 3449.0 | 10.5 | 70 | 1 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 393 | 27.0 | 4 | 140.0 | 86.0 | 2790.0 | 15.6 | 82 | 1 |
| 394 | 44.0 | 4 | 97.0 | 52.0 | 2130.0 | 24.6 | 82 | 2 |
| 395 | 32.0 | 4 | 135.0 | 84.0 | 2295.0 | 11.6 | 82 | 1 |
| 396 | 28.0 | 4 | 120.0 | 79.0 | 2625.0 | 18.6 | 82 | 1 |
| 397 | 31.0 | 4 | 119.0 | 82.0 | 2720.0 | 19.4 | 82 | 1 |
392 rows × 8 columns
dataset_np= dataset.to_numpy()
dataset_np
array([[ 18. , 8. , 307. , ..., 12. , 70. , 1. ],
[ 15. , 8. , 350. , ..., 11.5, 70. , 1. ],
[ 18. , 8. , 318. , ..., 11. , 70. , 1. ],
...,
[ 32. , 4. , 135. , ..., 11.6, 82. , 1. ],
[ 28. , 4. , 120. , ..., 18.6, 82. , 1. ],
[ 31. , 4. , 119. , ..., 19.4, 82. , 1. ]])
from sklearn.preprocessing import OneHotEncoder
# Son sütunu seç (kategori sütunu)
categorical_column = dataset_np[:, -1].reshape(-1, 1)
# One-hot encoder oluştur ve uygula
encoder = OneHotEncoder(sparse_output=False)
encoded_column = encoder.fit_transform(categorical_column)
# Sonucu orijinal veriyle birleştir
dataset_encoded = np.hstack((dataset_np[:, :-1], encoded_column))
print(dataset_encoded)
[[ 18. 8. 307. ... 1. 0. 0.]
[ 15. 8. 350. ... 1. 0. 0.]
[ 18. 8. 318. ... 1. 0. 0.]
...
[ 32. 4. 135. ... 1. 0. 0.]
[ 28. 4. 120. ... 1. 0. 0.]
[ 31. 4. 119. ... 1. 0. 0.]]
print(dataset_encoded[-5:,:])
[[2.700e+01 4.000e+00 1.400e+02 8.600e+01 2.790e+03 1.560e+01 8.200e+01
1.000e+00 0.000e+00 0.000e+00]
[4.400e+01 4.000e+00 9.700e+01 5.200e+01 2.130e+03 2.460e+01 8.200e+01
0.000e+00 1.000e+00 0.000e+00]
[3.200e+01 4.000e+00 1.350e+02 8.400e+01 2.295e+03 1.160e+01 8.200e+01
1.000e+00 0.000e+00 0.000e+00]
[2.800e+01 4.000e+00 1.200e+02 7.900e+01 2.625e+03 1.860e+01 8.200e+01
1.000e+00 0.000e+00 0.000e+00]
[3.100e+01 4.000e+00 1.190e+02 8.200e+01 2.720e+03 1.940e+01 8.200e+01
1.000e+00 0.000e+00 0.000e+00]]
# Horsepower
hp = dataset_encoded[:,3]
mpg = dataset_encoded[:,0]
plt.scatter(hp, mpg, marker="+", color="red")
plt.xlabel("hp")
plt.ylabel('mpg')
plt.show()
# Horsepower
weight = dataset_encoded[:,4]
mpg = dataset_encoded[:,0]
plt.scatter(weight, mpg, marker="+", color="blue")
plt.xlabel("weight")
plt.ylabel('mpg')
plt.show()
Tek katman, tek nöron
# input ve output
y = mpg # İlk sütun (çıktı)
X = hp # Geri kalan sütunlar (girdi)
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
from sklearn.preprocessing import MinMaxScaler
# X_train için Min-Max ölçekleme
scaler_X = MinMaxScaler()
# hp
X_train_scaled = scaler_X.fit_transform(X_train.reshape((-1,1)))
# X_test'i aynı scaler ile dönüştür
X_test_scaled = scaler_X.transform(X_test.reshape((-1,1)))
# y_train için Min-Max ölçekleme
scaler_y = MinMaxScaler()
y_train_scaled = scaler_y.fit_transform(y_train.reshape(-1, 1)) # y vektör olduğu için reshape gerekli
# y_test'i aynı scaler ile dönüştür
y_test_scaled = scaler_y.transform(y_test.reshape(-1, 1))
X_train_scaled.shape
(313, 1)
model = keras.models.Sequential([
layers.Dense(units=1, input_shape=(1,)) # Linear Model
])
model.summary()
Model: "sequential_10"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_12 (Dense) (None, 1) 2
=================================================================
Total params: 2
Trainable params: 2
Non-trainable params: 0
_________________________________________________________________
model.compile(optimizer='adam', loss='mse')
history=model.fit(X_train_scaled, y_train_scaled, epochs=100, validation_data=(X_test_scaled, y_test_scaled))
Epoch 1/100
10/10 [==============================] - 1s 23ms/step - loss: 0.1432 - val_loss: 0.1209
Epoch 2/100
10/10 [==============================] - 0s 10ms/step - loss: 0.1381 - val_loss: 0.1165
Epoch 3/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1334 - val_loss: 0.1124
Epoch 4/100
10/10 [==============================] - 0s 11ms/step - loss: 0.1291 - val_loss: 0.1088
Epoch 5/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1252 - val_loss: 0.1054
Epoch 6/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1215 - val_loss: 0.1025
Epoch 7/100
10/10 [==============================] - 0s 10ms/step - loss: 0.1183 - val_loss: 0.0998
Epoch 8/100
10/10 [==============================] - 0s 10ms/step - loss: 0.1154 - val_loss: 0.0972
Epoch 9/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1127 - val_loss: 0.0949
Epoch 10/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1102 - val_loss: 0.0927
Epoch 11/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1078 - val_loss: 0.0907
Epoch 12/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1056 - val_loss: 0.0888
Epoch 13/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1035 - val_loss: 0.0870
Epoch 14/100
10/10 [==============================] - 0s 9ms/step - loss: 0.1014 - val_loss: 0.0852
Epoch 15/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0995 - val_loss: 0.0835
Epoch 16/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0977 - val_loss: 0.0819
Epoch 17/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0958 - val_loss: 0.0803
Epoch 18/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0941 - val_loss: 0.0788
Epoch 19/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0925 - val_loss: 0.0773
Epoch 20/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0908 - val_loss: 0.0759
Epoch 21/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0892 - val_loss: 0.0745
Epoch 22/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0876 - val_loss: 0.0731
Epoch 23/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0862 - val_loss: 0.0717
Epoch 24/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0847 - val_loss: 0.0704
Epoch 25/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0832 - val_loss: 0.0692
Epoch 26/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0818 - val_loss: 0.0679
Epoch 27/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0805 - val_loss: 0.0667
Epoch 28/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0791 - val_loss: 0.0655
Epoch 29/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0779 - val_loss: 0.0644
Epoch 30/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0766 - val_loss: 0.0633
Epoch 31/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0754 - val_loss: 0.0621
Epoch 32/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0742 - val_loss: 0.0611
Epoch 33/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0730 - val_loss: 0.0600
Epoch 34/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0718 - val_loss: 0.0590
Epoch 35/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0707 - val_loss: 0.0580
Epoch 36/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0696 - val_loss: 0.0570
Epoch 37/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0685 - val_loss: 0.0560
Epoch 38/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0674 - val_loss: 0.0551
Epoch 39/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0664 - val_loss: 0.0542
Epoch 40/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0654 - val_loss: 0.0532
Epoch 41/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0644 - val_loss: 0.0523
Epoch 42/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0634 - val_loss: 0.0514
Epoch 43/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0624 - val_loss: 0.0506
Epoch 44/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0615 - val_loss: 0.0498
Epoch 45/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0605 - val_loss: 0.0490
Epoch 46/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0596 - val_loss: 0.0481
Epoch 47/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0588 - val_loss: 0.0473
Epoch 48/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0579 - val_loss: 0.0466
Epoch 49/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0570 - val_loss: 0.0458
Epoch 50/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0562 - val_loss: 0.0451
Epoch 51/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0554 - val_loss: 0.0444
Epoch 52/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0546 - val_loss: 0.0437
Epoch 53/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0538 - val_loss: 0.0430
Epoch 54/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0530 - val_loss: 0.0423
Epoch 55/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0523 - val_loss: 0.0416
Epoch 56/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0515 - val_loss: 0.0409
Epoch 57/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0508 - val_loss: 0.0403
Epoch 58/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0501 - val_loss: 0.0397
Epoch 59/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0493 - val_loss: 0.0390
Epoch 60/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0486 - val_loss: 0.0384
Epoch 61/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0480 - val_loss: 0.0379
Epoch 62/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0473 - val_loss: 0.0373
Epoch 63/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0467 - val_loss: 0.0367
Epoch 64/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0460 - val_loss: 0.0361
Epoch 65/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0454 - val_loss: 0.0356
Epoch 66/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0448 - val_loss: 0.0351
Epoch 67/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0442 - val_loss: 0.0346
Epoch 68/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0436 - val_loss: 0.0341
Epoch 69/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0430 - val_loss: 0.0336
Epoch 70/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0424 - val_loss: 0.0331
Epoch 71/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0419 - val_loss: 0.0326
Epoch 72/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0413 - val_loss: 0.0321
Epoch 73/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0408 - val_loss: 0.0317
Epoch 74/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0403 - val_loss: 0.0312
Epoch 75/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0397 - val_loss: 0.0308
Epoch 76/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0393 - val_loss: 0.0303
Epoch 77/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0387 - val_loss: 0.0299
Epoch 78/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0383 - val_loss: 0.0295
Epoch 79/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0378 - val_loss: 0.0291
Epoch 80/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0373 - val_loss: 0.0287
Epoch 81/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0368 - val_loss: 0.0283
Epoch 82/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0364 - val_loss: 0.0279
Epoch 83/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0359 - val_loss: 0.0276
Epoch 84/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0355 - val_loss: 0.0272
Epoch 85/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0351 - val_loss: 0.0268
Epoch 86/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0347 - val_loss: 0.0265
Epoch 87/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0343 - val_loss: 0.0261
Epoch 88/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0338 - val_loss: 0.0258
Epoch 89/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0334 - val_loss: 0.0255
Epoch 90/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0331 - val_loss: 0.0252
Epoch 91/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0327 - val_loss: 0.0249
Epoch 92/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0323 - val_loss: 0.0246
Epoch 93/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0319 - val_loss: 0.0243
Epoch 94/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0316 - val_loss: 0.0240
Epoch 95/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0312 - val_loss: 0.0237
Epoch 96/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0309 - val_loss: 0.0234
Epoch 97/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0306 - val_loss: 0.0232
Epoch 98/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0302 - val_loss: 0.0229
Epoch 99/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0299 - val_loss: 0.0226
Epoch 100/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0296 - val_loss: 0.0224
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Error [MPG]')
plt.legend()
plt.grid(True)
plt.show()
y_test_predict_scaled=model.predict(X_test_scaled)
3/3 [==============================] - 0s 3ms/step
y_pred = scaler_y.inverse_transform(y_test_predict_scaled)
y_test[:10]
array([26. , 21.6, 36.1, 26. , 27. , 28. , 13. , 26. , 19. , 29. ])
y_pred[:10]
array([[24.879032],
[22.3061 ],
[25.382433],
[24.8231 ],
[23.928167],
[24.543432],
[19.229767],
[24.543432],
[23.424765],
[25.997698]], dtype=float32)
X_fake = np.linspace(50, 230, 200).reshape((-1,1))
X_fake_scaled = scaler_X.transform(X_fake)
y_fake_pred_scaled = model.predict(X_fake_scaled)
y_fake_pred=scaler_y.inverse_transform(y_fake_pred_scaled)
7/7 [==============================] - 0s 3ms/step
plt.plot(X_fake, y_fake_pred)
plt.scatter(hp, mpg, marker="+", color="red")
plt.show()
Aktivasyonsuz büyük model
# DNN
model = keras.Sequential([
layers.Dense(64, input_shape=(1,)),
layers.Dense(64),
layers.Dense(1)
])
model.compile(loss="mse", optimizer="adam")
model.summary()
Model: "sequential_11"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_13 (Dense) (None, 64) 128
dense_14 (Dense) (None, 64) 4160
dense_15 (Dense) (None, 1) 65
=================================================================
Total params: 4,353
Trainable params: 4,353
Non-trainable params: 0
_________________________________________________________________
history=model.fit(X_train_scaled, y_train_scaled, epochs=100, validation_data=(X_test_scaled, y_test_scaled))
Epoch 1/100
10/10 [==============================] - 1s 21ms/step - loss: 0.1110 - val_loss: 0.0613
Epoch 2/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0525 - val_loss: 0.0222
Epoch 3/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0214 - val_loss: 0.0166
Epoch 4/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0191 - val_loss: 0.0210
Epoch 5/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0196 - val_loss: 0.0173
Epoch 6/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0194 - val_loss: 0.0171
Epoch 7/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0174 - val_loss: 0.0152
Epoch 8/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0176 - val_loss: 0.0160
Epoch 9/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0174 - val_loss: 0.0158
Epoch 10/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0161
Epoch 11/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0185 - val_loss: 0.0182
Epoch 12/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0187 - val_loss: 0.0153
Epoch 13/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0191 - val_loss: 0.0154
Epoch 14/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0181 - val_loss: 0.0158
Epoch 15/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0189 - val_loss: 0.0186
Epoch 16/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0186 - val_loss: 0.0163
Epoch 17/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0182 - val_loss: 0.0153
Epoch 18/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0181 - val_loss: 0.0156
Epoch 19/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0163
Epoch 20/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0183 - val_loss: 0.0162
Epoch 21/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0156
Epoch 22/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0184 - val_loss: 0.0160
Epoch 23/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0185 - val_loss: 0.0169
Epoch 24/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0155
Epoch 25/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0174 - val_loss: 0.0158
Epoch 26/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0174 - val_loss: 0.0156
Epoch 27/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0182 - val_loss: 0.0176
Epoch 28/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0158
Epoch 29/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0157
Epoch 30/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0156
Epoch 31/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0174 - val_loss: 0.0155
Epoch 32/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0184 - val_loss: 0.0155
Epoch 33/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0181 - val_loss: 0.0158
Epoch 34/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0163
Epoch 35/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0181 - val_loss: 0.0175
Epoch 36/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0188 - val_loss: 0.0162
Epoch 37/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0194 - val_loss: 0.0156
Epoch 38/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0175 - val_loss: 0.0162
Epoch 39/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0179 - val_loss: 0.0158
Epoch 40/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0190 - val_loss: 0.0159
Epoch 41/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0188 - val_loss: 0.0159
Epoch 42/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0163
Epoch 43/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0154
Epoch 44/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0179 - val_loss: 0.0166
Epoch 45/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0158
Epoch 46/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0174 - val_loss: 0.0158
Epoch 47/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0167
Epoch 48/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0179 - val_loss: 0.0153
Epoch 49/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0153
Epoch 50/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0159
Epoch 51/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0167
Epoch 52/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0178 - val_loss: 0.0155
Epoch 53/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0161
Epoch 54/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0177 - val_loss: 0.0167
Epoch 55/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0182 - val_loss: 0.0163
Epoch 56/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0154
Epoch 57/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0155
Epoch 58/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0155
Epoch 59/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0163
Epoch 60/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0181 - val_loss: 0.0159
Epoch 61/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0180 - val_loss: 0.0156
Epoch 62/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0176 - val_loss: 0.0174
Epoch 63/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0155
Epoch 64/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0176 - val_loss: 0.0160
Epoch 65/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0155
Epoch 66/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0175 - val_loss: 0.0161
Epoch 67/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0176 - val_loss: 0.0167
Epoch 68/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0182 - val_loss: 0.0159
Epoch 69/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0186 - val_loss: 0.0155
Epoch 70/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0183 - val_loss: 0.0154
Epoch 71/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0164
Epoch 72/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0180 - val_loss: 0.0159
Epoch 73/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0174 - val_loss: 0.0158
Epoch 74/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0174 - val_loss: 0.0157
Epoch 75/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0177 - val_loss: 0.0158
Epoch 76/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0161
Epoch 77/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0156
Epoch 78/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0176 - val_loss: 0.0156
Epoch 79/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0176 - val_loss: 0.0166
Epoch 80/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0155
Epoch 81/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0175 - val_loss: 0.0163
Epoch 82/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0179 - val_loss: 0.0158
Epoch 83/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0160
Epoch 84/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0175 - val_loss: 0.0153
Epoch 85/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0161
Epoch 86/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0180 - val_loss: 0.0178
Epoch 87/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0179 - val_loss: 0.0153
Epoch 88/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0177 - val_loss: 0.0158
Epoch 89/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0176 - val_loss: 0.0160
Epoch 90/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0177 - val_loss: 0.0161
Epoch 91/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0156
Epoch 92/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0181 - val_loss: 0.0161
Epoch 93/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0172
Epoch 94/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0180 - val_loss: 0.0150
Epoch 95/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0178 - val_loss: 0.0156
Epoch 96/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0176 - val_loss: 0.0165
Epoch 97/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0177 - val_loss: 0.0165
Epoch 98/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0181 - val_loss: 0.0156
Epoch 99/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0179 - val_loss: 0.0153
Epoch 100/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0182 - val_loss: 0.0156
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Error [MPG]')
plt.legend()
plt.grid(True)
plt.show()
y_test_predict_scaled=model.predict(X_test_scaled)
3/3 [==============================] - 0s 3ms/step
y_pred = scaler_y.inverse_transform(y_test_predict_scaled)
y_test[:10]
array([26. , 21.6, 36.1, 26. , 27. , 28. , 13. , 26. , 19. , 29. ])
y_pred[:10]
array([[29.274797],
[22.273424],
[30.642967],
[29.121162],
[26.68683 ],
[28.362114],
[13.90795 ],
[28.362114],
[25.316181],
[32.31797 ]], dtype=float32)
X_fake = np.linspace(50, 230, 200).reshape((-1,1))
X_fake_scaled = scaler_X.transform(X_fake)
y_fake_pred_scaled = model.predict(X_fake_scaled)
y_fake_pred=scaler_y.inverse_transform(y_fake_pred_scaled)
7/7 [==============================] - 0s 2ms/step
plt.plot(X_fake, y_fake_pred)
plt.scatter(hp, mpg, marker="+", color="red")
plt.show()
Aktivasyonlu büyük model
# DNN
model = keras.Sequential([
layers.Dense(64, activation='relu', input_shape=(1,)),
layers.Dense(64, activation='relu'),
layers.Dense(1)
])
model.compile(loss="mse", optimizer="adam")
model.summary()
Model: "sequential_12"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_16 (Dense) (None, 64) 128
dense_17 (Dense) (None, 64) 4160
dense_18 (Dense) (None, 1) 65
=================================================================
Total params: 4,353
Trainable params: 4,353
Non-trainable params: 0
_________________________________________________________________
history=model.fit(X_train_scaled, y_train_scaled, epochs=100, validation_data=(X_test_scaled, y_test_scaled))
Epoch 1/100
10/10 [==============================] - 1s 22ms/step - loss: 0.1405 - val_loss: 0.0941
Epoch 2/100
10/10 [==============================] - 0s 10ms/step - loss: 0.1041 - val_loss: 0.0799
Epoch 3/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0858 - val_loss: 0.0634
Epoch 4/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0697 - val_loss: 0.0500
Epoch 5/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0562 - val_loss: 0.0383
Epoch 6/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0436 - val_loss: 0.0288
Epoch 7/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0330 - val_loss: 0.0211
Epoch 8/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0251 - val_loss: 0.0164
Epoch 9/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0200 - val_loss: 0.0147
Epoch 10/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0179 - val_loss: 0.0143
Epoch 11/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0168 - val_loss: 0.0144
Epoch 12/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0165 - val_loss: 0.0144
Epoch 13/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0161 - val_loss: 0.0134
Epoch 14/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0160 - val_loss: 0.0138
Epoch 15/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0155 - val_loss: 0.0129
Epoch 16/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0152 - val_loss: 0.0133
Epoch 17/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0151 - val_loss: 0.0131
Epoch 18/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0151 - val_loss: 0.0135
Epoch 19/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0148 - val_loss: 0.0128
Epoch 20/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0146 - val_loss: 0.0129
Epoch 21/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0144 - val_loss: 0.0133
Epoch 22/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0143 - val_loss: 0.0130
Epoch 23/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0143 - val_loss: 0.0134
Epoch 24/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0143 - val_loss: 0.0127
Epoch 25/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0142 - val_loss: 0.0141
Epoch 26/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0145 - val_loss: 0.0127
Epoch 27/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0143 - val_loss: 0.0136
Epoch 28/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0140 - val_loss: 0.0126
Epoch 29/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0142 - val_loss: 0.0141
Epoch 30/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0140 - val_loss: 0.0126
Epoch 31/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0139 - val_loss: 0.0133
Epoch 32/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0139 - val_loss: 0.0126
Epoch 33/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0138 - val_loss: 0.0132
Epoch 34/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0138 - val_loss: 0.0129
Epoch 35/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0141 - val_loss: 0.0124
Epoch 36/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0138 - val_loss: 0.0136
Epoch 37/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0137 - val_loss: 0.0125
Epoch 38/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0137 - val_loss: 0.0128
Epoch 39/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0136 - val_loss: 0.0131
Epoch 40/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0136 - val_loss: 0.0126
Epoch 41/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0138 - val_loss: 0.0132
Epoch 42/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0135 - val_loss: 0.0126
Epoch 43/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0136 - val_loss: 0.0133
Epoch 44/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0134 - val_loss: 0.0125
Epoch 45/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0135 - val_loss: 0.0125
Epoch 46/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0134 - val_loss: 0.0133
Epoch 47/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0135 - val_loss: 0.0126
Epoch 48/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0134 - val_loss: 0.0128
Epoch 49/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0134 - val_loss: 0.0127
Epoch 50/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0135 - val_loss: 0.0126
Epoch 51/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0135 - val_loss: 0.0131
Epoch 52/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0135 - val_loss: 0.0123
Epoch 53/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0134 - val_loss: 0.0130
Epoch 54/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0133 - val_loss: 0.0125
Epoch 55/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0128
Epoch 56/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0135 - val_loss: 0.0131
Epoch 57/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0141 - val_loss: 0.0134
Epoch 58/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0133 - val_loss: 0.0122
Epoch 59/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0134 - val_loss: 0.0128
Epoch 60/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0128
Epoch 61/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0133 - val_loss: 0.0129
Epoch 62/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0132 - val_loss: 0.0128
Epoch 63/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0132
Epoch 64/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0134 - val_loss: 0.0126
Epoch 65/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0125
Epoch 66/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0132 - val_loss: 0.0127
Epoch 67/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0132 - val_loss: 0.0129
Epoch 68/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0135 - val_loss: 0.0125
Epoch 69/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0134 - val_loss: 0.0129
Epoch 70/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0123
Epoch 71/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0132
Epoch 72/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0124
Epoch 73/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0133 - val_loss: 0.0126
Epoch 74/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0134 - val_loss: 0.0125
Epoch 75/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0129
Epoch 76/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0136 - val_loss: 0.0133
Epoch 77/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0135 - val_loss: 0.0126
Epoch 78/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0130
Epoch 79/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0131 - val_loss: 0.0123
Epoch 80/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0134 - val_loss: 0.0134
Epoch 81/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0132 - val_loss: 0.0126
Epoch 82/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0133 - val_loss: 0.0130
Epoch 83/100
10/10 [==============================] - 0s 12ms/step - loss: 0.0132 - val_loss: 0.0126
Epoch 84/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0132 - val_loss: 0.0127
Epoch 85/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0132 - val_loss: 0.0128
Epoch 86/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0133 - val_loss: 0.0128
Epoch 87/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0127
Epoch 88/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0132 - val_loss: 0.0127
Epoch 89/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0133 - val_loss: 0.0125
Epoch 90/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0134 - val_loss: 0.0135
Epoch 91/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0137 - val_loss: 0.0124
Epoch 92/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0133 - val_loss: 0.0131
Epoch 93/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0133 - val_loss: 0.0127
Epoch 94/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0132 - val_loss: 0.0129
Epoch 95/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0132 - val_loss: 0.0125
Epoch 96/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0132 - val_loss: 0.0128
Epoch 97/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0133 - val_loss: 0.0135
Epoch 98/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0134 - val_loss: 0.0125
Epoch 99/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0132 - val_loss: 0.0134
Epoch 100/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0133 - val_loss: 0.0129
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Error [MPG]')
plt.legend()
plt.grid(True)
plt.show()
y_test_predict_scaled=model.predict(X_test_scaled)
3/3 [==============================] - 0s 3ms/step
y_pred = scaler_y.inverse_transform(y_test_predict_scaled)
y_test[:10]
array([26. , 21.6, 36.1, 26. , 27. , 28. , 13. , 26. , 19. , 29. ])
y_pred[:10]
array([[32.31129 ],
[19.190641],
[34.19633 ],
[31.933548],
[25.983114],
[30.077898],
[14.251822],
[30.077898],
[22.63136 ],
[35.136513]], dtype=float32)
X_fake = np.linspace(50, 230, 200).reshape((-1,1))
X_fake_scaled = scaler_X.transform(X_fake)
y_fake_pred_scaled = model.predict(X_fake_scaled)
y_fake_pred=scaler_y.inverse_transform(y_fake_pred_scaled)
7/7 [==============================] - 0s 3ms/step
plt.plot(X_fake, y_fake_pred)
plt.scatter(hp, mpg, marker="+", color="red")
plt.show()
Çoklu veri
X = dataset_encoded[:,1:]
y = dataset_encoded[:,0]
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
from sklearn.preprocessing import MinMaxScaler
# X_train için Min-Max ölçekleme
scaler_X = MinMaxScaler()
# hp
X_train_scaled = scaler_X.fit_transform(X_train)
# X_test'i aynı scaler ile dönüştür
X_test_scaled = scaler_X.transform(X_test)
# y_train için Min-Max ölçekleme
scaler_y = MinMaxScaler()
y_train_scaled = scaler_y.fit_transform(y_train.reshape(-1, 1)) # y vektör olduğu için reshape gerekli
# y_test'i aynı scaler ile dönüştür
y_test_scaled = scaler_y.transform(y_test.reshape(-1, 1))
X_train_scaled.shape, X_test_scaled.shape, y_train.shape
((313, 9), (79, 9), (313,))
Aktivasyon fonksiyonu olmayan model
model = keras.models.Sequential([
layers.Dense(units=32, input_shape=(9,)), # Linear Model
layers.Dense(1)
])
model.summary()
Model: "sequential_13"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_19 (Dense) (None, 32) 320
dense_20 (Dense) (None, 1) 33
=================================================================
Total params: 353
Trainable params: 353
Non-trainable params: 0
_________________________________________________________________
model.compile(optimizer='adam', loss='mse')
history=model.fit(X_train_scaled, y_train_scaled, epochs=100, validation_data=(X_test_scaled, y_test_scaled))
Epoch 1/100
10/10 [==============================] - 1s 25ms/step - loss: 0.1336 - val_loss: 0.0811
Epoch 2/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0663 - val_loss: 0.0381
Epoch 3/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0298 - val_loss: 0.0159
Epoch 4/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0153 - val_loss: 0.0110
Epoch 5/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0115 - val_loss: 0.0114
Epoch 6/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0115 - val_loss: 0.0119
Epoch 7/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0112 - val_loss: 0.0113
Epoch 8/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0109 - val_loss: 0.0106
Epoch 9/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0108 - val_loss: 0.0101
Epoch 10/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0104 - val_loss: 0.0099
Epoch 11/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0103 - val_loss: 0.0098
Epoch 12/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0102 - val_loss: 0.0098
Epoch 13/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0100 - val_loss: 0.0097
Epoch 14/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0098 - val_loss: 0.0095
Epoch 15/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0097 - val_loss: 0.0094
Epoch 16/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0096 - val_loss: 0.0093
Epoch 17/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0095 - val_loss: 0.0091
Epoch 18/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0093 - val_loss: 0.0093
Epoch 19/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0092 - val_loss: 0.0093
Epoch 20/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0092 - val_loss: 0.0092
Epoch 21/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0090 - val_loss: 0.0088
Epoch 22/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0090 - val_loss: 0.0089
Epoch 23/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0089 - val_loss: 0.0088
Epoch 24/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0089 - val_loss: 0.0088
Epoch 25/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0088 - val_loss: 0.0087
Epoch 26/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0087 - val_loss: 0.0086
Epoch 27/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0086 - val_loss: 0.0087
Epoch 28/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0085 - val_loss: 0.0086
Epoch 29/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0084 - val_loss: 0.0085
Epoch 30/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0085 - val_loss: 0.0086
Epoch 31/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0084 - val_loss: 0.0086
Epoch 32/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0084 - val_loss: 0.0083
Epoch 33/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0083 - val_loss: 0.0085
Epoch 34/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0083 - val_loss: 0.0083
Epoch 35/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0084 - val_loss: 0.0082
Epoch 36/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0083 - val_loss: 0.0085
Epoch 37/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0084 - val_loss: 0.0084
Epoch 38/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0083 - val_loss: 0.0084
Epoch 39/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0082 - val_loss: 0.0084
Epoch 40/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0082 - val_loss: 0.0082
Epoch 41/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0082 - val_loss: 0.0083
Epoch 42/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0081 - val_loss: 0.0080
Epoch 43/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0085 - val_loss: 0.0081
Epoch 44/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0083 - val_loss: 0.0086
Epoch 45/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0082 - val_loss: 0.0085
Epoch 46/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0080 - val_loss: 0.0081
Epoch 47/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0081 - val_loss: 0.0079
Epoch 48/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0081 - val_loss: 0.0083
Epoch 49/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0081 - val_loss: 0.0081
Epoch 50/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0081 - val_loss: 0.0081
Epoch 51/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0081 - val_loss: 0.0081
Epoch 52/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0080 - val_loss: 0.0081
Epoch 53/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0080 - val_loss: 0.0083
Epoch 54/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0080 - val_loss: 0.0076
Epoch 55/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0082
Epoch 56/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0081
Epoch 57/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0081 - val_loss: 0.0080
Epoch 58/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0079 - val_loss: 0.0080
Epoch 59/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0080 - val_loss: 0.0079
Epoch 60/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0080 - val_loss: 0.0079
Epoch 61/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0080
Epoch 62/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0079 - val_loss: 0.0081
Epoch 63/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0079 - val_loss: 0.0081
Epoch 64/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0080 - val_loss: 0.0080
Epoch 65/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0081 - val_loss: 0.0082
Epoch 66/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0084 - val_loss: 0.0083
Epoch 67/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0082 - val_loss: 0.0087
Epoch 68/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0081
Epoch 69/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0080
Epoch 70/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0080 - val_loss: 0.0076
Epoch 71/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0081 - val_loss: 0.0082
Epoch 72/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0081
Epoch 73/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0078 - val_loss: 0.0080
Epoch 74/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0078 - val_loss: 0.0079
Epoch 75/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0077 - val_loss: 0.0078
Epoch 76/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0078 - val_loss: 0.0077
Epoch 77/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0078 - val_loss: 0.0079
Epoch 78/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0079 - val_loss: 0.0081
Epoch 79/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0078 - val_loss: 0.0077
Epoch 80/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0078 - val_loss: 0.0078
Epoch 81/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0079 - val_loss: 0.0083
Epoch 82/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0081 - val_loss: 0.0081
Epoch 83/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0081 - val_loss: 0.0081
Epoch 84/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0079 - val_loss: 0.0075
Epoch 85/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0079 - val_loss: 0.0080
Epoch 86/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0078 - val_loss: 0.0077
Epoch 87/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0078 - val_loss: 0.0082
Epoch 88/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0075
Epoch 89/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0078 - val_loss: 0.0080
Epoch 90/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0079 - val_loss: 0.0080
Epoch 91/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0082 - val_loss: 0.0081
Epoch 92/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0081 - val_loss: 0.0079
Epoch 93/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0078 - val_loss: 0.0076
Epoch 94/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0078 - val_loss: 0.0078
Epoch 95/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0078 - val_loss: 0.0080
Epoch 96/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0079 - val_loss: 0.0082
Epoch 97/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0079 - val_loss: 0.0077
Epoch 98/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0080 - val_loss: 0.0078
Epoch 99/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0079 - val_loss: 0.0079
Epoch 100/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0080 - val_loss: 0.0079
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Error [MPG]')
plt.legend()
plt.grid(True)
plt.show()
Aktivasyonlu tek katman model
model = keras.models.Sequential([
layers.Dense(units=32, input_shape=(9,), activation='relu'),
layers.Dense(1)
])
model.summary()
Model: "sequential_15"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_23 (Dense) (None, 32) 320
dense_24 (Dense) (None, 1) 33
=================================================================
Total params: 353
Trainable params: 353
Non-trainable params: 0
_________________________________________________________________
model.compile(optimizer='adam', loss='mse')
history=model.fit(X_train_scaled, y_train_scaled, epochs=100, validation_data=(X_test_scaled, y_test_scaled))
Epoch 1/100
10/10 [==============================] - 1s 21ms/step - loss: 0.0943 - val_loss: 0.0820
Epoch 2/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0606 - val_loss: 0.0554
Epoch 3/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0393 - val_loss: 0.0382
Epoch 4/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0266 - val_loss: 0.0273
Epoch 5/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0193 - val_loss: 0.0207
Epoch 6/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0151 - val_loss: 0.0166
Epoch 7/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0126 - val_loss: 0.0141
Epoch 8/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0114 - val_loss: 0.0127
Epoch 9/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0106 - val_loss: 0.0120
Epoch 10/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0103 - val_loss: 0.0115
Epoch 11/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0100 - val_loss: 0.0113
Epoch 12/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0097 - val_loss: 0.0111
Epoch 13/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0095 - val_loss: 0.0108
Epoch 14/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0093 - val_loss: 0.0107
Epoch 15/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0092 - val_loss: 0.0103
Epoch 16/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0089 - val_loss: 0.0102
Epoch 17/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0088 - val_loss: 0.0100
Epoch 18/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0086 - val_loss: 0.0101
Epoch 19/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0085 - val_loss: 0.0097
Epoch 20/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0083 - val_loss: 0.0096
Epoch 21/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0081 - val_loss: 0.0094
Epoch 22/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0081 - val_loss: 0.0092
Epoch 23/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0080 - val_loss: 0.0092
Epoch 24/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0078 - val_loss: 0.0089
Epoch 25/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0077 - val_loss: 0.0090
Epoch 26/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0076 - val_loss: 0.0087
Epoch 27/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0074 - val_loss: 0.0085
Epoch 28/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0074 - val_loss: 0.0084
Epoch 29/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0073 - val_loss: 0.0081
Epoch 30/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0070 - val_loss: 0.0082
Epoch 31/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0070 - val_loss: 0.0080
Epoch 32/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0069 - val_loss: 0.0079
Epoch 33/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0067 - val_loss: 0.0078
Epoch 34/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0067 - val_loss: 0.0076
Epoch 35/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0066 - val_loss: 0.0076
Epoch 36/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0066 - val_loss: 0.0075
Epoch 37/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0064 - val_loss: 0.0073
Epoch 38/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0063 - val_loss: 0.0072
Epoch 39/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0063 - val_loss: 0.0070
Epoch 40/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0062 - val_loss: 0.0070
Epoch 41/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0061 - val_loss: 0.0070
Epoch 42/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0061 - val_loss: 0.0070
Epoch 43/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0060 - val_loss: 0.0067
Epoch 44/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0059 - val_loss: 0.0068
Epoch 45/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0060 - val_loss: 0.0069
Epoch 46/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0059 - val_loss: 0.0067
Epoch 47/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0058 - val_loss: 0.0066
Epoch 48/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0057 - val_loss: 0.0064
Epoch 49/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0057 - val_loss: 0.0064
Epoch 50/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0057 - val_loss: 0.0065
Epoch 51/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0056 - val_loss: 0.0062
Epoch 52/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0056 - val_loss: 0.0062
Epoch 53/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0055 - val_loss: 0.0062
Epoch 54/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0055 - val_loss: 0.0061
Epoch 55/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0055 - val_loss: 0.0060
Epoch 56/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0054 - val_loss: 0.0061
Epoch 57/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0054 - val_loss: 0.0060
Epoch 58/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0053 - val_loss: 0.0060
Epoch 59/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0053 - val_loss: 0.0059
Epoch 60/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0053 - val_loss: 0.0058
Epoch 61/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0052 - val_loss: 0.0058
Epoch 62/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0052 - val_loss: 0.0058
Epoch 63/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0053 - val_loss: 0.0058
Epoch 64/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0052 - val_loss: 0.0057
Epoch 65/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0052 - val_loss: 0.0058
Epoch 66/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0052 - val_loss: 0.0057
Epoch 67/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0051 - val_loss: 0.0057
Epoch 68/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0051 - val_loss: 0.0055
Epoch 69/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0051 - val_loss: 0.0057
Epoch 70/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0051 - val_loss: 0.0057
Epoch 71/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0051 - val_loss: 0.0056
Epoch 72/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0051 - val_loss: 0.0055
Epoch 73/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0056
Epoch 74/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0056
Epoch 75/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0051 - val_loss: 0.0054
Epoch 76/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0051 - val_loss: 0.0055
Epoch 77/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0051 - val_loss: 0.0055
Epoch 78/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0054
Epoch 79/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0055
Epoch 80/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0050 - val_loss: 0.0055
Epoch 81/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0054
Epoch 82/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0049 - val_loss: 0.0054
Epoch 83/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0049 - val_loss: 0.0053
Epoch 84/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0050 - val_loss: 0.0055
Epoch 85/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0054
Epoch 86/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0053
Epoch 87/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0050 - val_loss: 0.0054
Epoch 88/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0054
Epoch 89/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0049 - val_loss: 0.0054
Epoch 90/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0053
Epoch 91/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0053
Epoch 92/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0050 - val_loss: 0.0053
Epoch 93/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0049 - val_loss: 0.0053
Epoch 94/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0049 - val_loss: 0.0055
Epoch 95/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0049 - val_loss: 0.0053
Epoch 96/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0048 - val_loss: 0.0052
Epoch 97/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0048 - val_loss: 0.0052
Epoch 98/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0048 - val_loss: 0.0052
Epoch 99/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0048 - val_loss: 0.0052
Epoch 100/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0048 - val_loss: 0.0052
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Error [MPG]')
plt.legend()
plt.grid(True)
plt.show()
Aktivasyonlu büyük model
model = keras.models.Sequential([
layers.Dense(units=64, input_shape=(9,), activation='relu'),
layers.Dense(units=64, input_shape=(9,), activation='relu'),
layers.Dense(1)
])
model.summary()
Model: "sequential_16"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_25 (Dense) (None, 64) 640
dense_26 (Dense) (None, 64) 4160
dense_27 (Dense) (None, 1) 65
=================================================================
Total params: 4,865
Trainable params: 4,865
Non-trainable params: 0
_________________________________________________________________
model.compile(optimizer='adam', loss='mse')
history=model.fit(X_train_scaled, y_train_scaled, epochs=100, validation_data=(X_test_scaled, y_test_scaled))
Epoch 1/100
10/10 [==============================] - 1s 22ms/step - loss: 0.1527 - val_loss: 0.0731
Epoch 2/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0637 - val_loss: 0.0294
Epoch 3/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0233 - val_loss: 0.0108
Epoch 4/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0116 - val_loss: 0.0123
Epoch 5/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0097 - val_loss: 0.0094
Epoch 6/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0081 - val_loss: 0.0078
Epoch 7/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0074 - val_loss: 0.0071
Epoch 8/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0070 - val_loss: 0.0071
Epoch 9/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0067 - val_loss: 0.0063
Epoch 10/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0064 - val_loss: 0.0063
Epoch 11/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0061 - val_loss: 0.0058
Epoch 12/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0059 - val_loss: 0.0057
Epoch 13/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0057 - val_loss: 0.0057
Epoch 14/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0055 - val_loss: 0.0057
Epoch 15/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0053 - val_loss: 0.0057
Epoch 16/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0053 - val_loss: 0.0055
Epoch 17/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0052 - val_loss: 0.0051
Epoch 18/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0051 - val_loss: 0.0057
Epoch 19/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0050 - val_loss: 0.0050
Epoch 20/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0051 - val_loss: 0.0052
Epoch 21/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0049 - val_loss: 0.0048
Epoch 22/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0048 - val_loss: 0.0055
Epoch 23/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0047 - val_loss: 0.0051
Epoch 24/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0046 - val_loss: 0.0049
Epoch 25/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0047 - val_loss: 0.0048
Epoch 26/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0046 - val_loss: 0.0050
Epoch 27/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0046 - val_loss: 0.0049
Epoch 28/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0048 - val_loss: 0.0051
Epoch 29/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0045 - val_loss: 0.0046
Epoch 30/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0046 - val_loss: 0.0051
Epoch 31/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0046 - val_loss: 0.0048
Epoch 32/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0045 - val_loss: 0.0048
Epoch 33/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0045 - val_loss: 0.0051
Epoch 34/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0044 - val_loss: 0.0048
Epoch 35/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0043 - val_loss: 0.0047
Epoch 36/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0042 - val_loss: 0.0046
Epoch 37/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0043 - val_loss: 0.0045
Epoch 38/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0044
Epoch 39/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0048
Epoch 40/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0045
Epoch 41/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0043 - val_loss: 0.0048
Epoch 42/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0041 - val_loss: 0.0044
Epoch 43/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0042 - val_loss: 0.0046
Epoch 44/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0044
Epoch 45/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0041 - val_loss: 0.0045
Epoch 46/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0042 - val_loss: 0.0044
Epoch 47/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0043
Epoch 48/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0041 - val_loss: 0.0046
Epoch 49/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0043
Epoch 50/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0040 - val_loss: 0.0044
Epoch 51/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0045
Epoch 52/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0041 - val_loss: 0.0048
Epoch 53/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0042 - val_loss: 0.0044
Epoch 54/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0041 - val_loss: 0.0040
Epoch 55/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0040 - val_loss: 0.0043
Epoch 56/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0044
Epoch 57/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0039 - val_loss: 0.0041
Epoch 58/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0039 - val_loss: 0.0044
Epoch 59/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0039 - val_loss: 0.0046
Epoch 60/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0039 - val_loss: 0.0043
Epoch 61/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0038 - val_loss: 0.0040
Epoch 62/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0047
Epoch 63/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0039 - val_loss: 0.0040
Epoch 64/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0039 - val_loss: 0.0044
Epoch 65/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0038 - val_loss: 0.0040
Epoch 66/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0038 - val_loss: 0.0041
Epoch 67/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0038 - val_loss: 0.0045
Epoch 68/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0043
Epoch 69/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0039 - val_loss: 0.0042
Epoch 70/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0038 - val_loss: 0.0040
Epoch 71/100
10/10 [==============================] - 0s 10ms/step - loss: 0.0039 - val_loss: 0.0050
Epoch 72/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0044
Epoch 73/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0039 - val_loss: 0.0039
Epoch 74/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0037 - val_loss: 0.0042
Epoch 75/100
10/10 [==============================] - 0s 11ms/step - loss: 0.0038 - val_loss: 0.0041
Epoch 76/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0037 - val_loss: 0.0041
Epoch 77/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0036 - val_loss: 0.0042
Epoch 78/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0037 - val_loss: 0.0040
Epoch 79/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0036 - val_loss: 0.0044
Epoch 80/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0038 - val_loss: 0.0043
Epoch 81/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0048
Epoch 82/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0039 - val_loss: 0.0042
Epoch 83/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0042 - val_loss: 0.0043
Epoch 84/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0043
Epoch 85/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0036 - val_loss: 0.0041
Epoch 86/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0036 - val_loss: 0.0045
Epoch 87/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0038 - val_loss: 0.0040
Epoch 88/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0037 - val_loss: 0.0042
Epoch 89/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0037 - val_loss: 0.0044
Epoch 90/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0038 - val_loss: 0.0044
Epoch 91/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0036 - val_loss: 0.0041
Epoch 92/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0037 - val_loss: 0.0041
Epoch 93/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0034 - val_loss: 0.0040
Epoch 94/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0035 - val_loss: 0.0044
Epoch 95/100
10/10 [==============================] - 0s 8ms/step - loss: 0.0037 - val_loss: 0.0042
Epoch 96/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0038 - val_loss: 0.0042
Epoch 97/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0040 - val_loss: 0.0044
Epoch 98/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0036 - val_loss: 0.0043
Epoch 99/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0035 - val_loss: 0.0043
Epoch 100/100
10/10 [==============================] - 0s 9ms/step - loss: 0.0036 - val_loss: 0.0039
plt.plot(history.history['loss'], label='loss')
plt.plot(history.history['val_loss'], label='val_loss')
plt.xlabel('Epoch')
plt.ylabel('Error [MPG]')
plt.legend()
plt.grid(True)
plt.show()
y_test_predict_scaled=model.predict(X_test_scaled)
3/3 [==============================] - 0s 4ms/step
y_pred = scaler_y.inverse_transform(y_test_predict_scaled)
y_test[:10]
array([26. , 21.6, 36.1, 26. , 27. , 28. , 13. , 26. , 19. , 29. ])
y_pred[:10]
array([[25.561352],
[21.317728],
[34.021828],
[21.4228 ],
[27.892326],
[28.932383],
[12.169561],
[29.77778 ],
[18.302343],
[30.233294]], dtype=float32)