Series temporales: temperaturas en Melbourne

• Introducción

Introducción

Se analizará la serie temporal de temperaturas mínimas en Melbourne con el dataset disponible en MachineLearningMastery.com de Jason Brownlee.

import tensorflow as tf
print(tf.__version__)

import numpy as np
import matplotlib.pyplot as plt
def plot_series(time, series, format="-", start=0, end=None):
    plt.plot(time[start:end], series[start:end], format)
    plt.xlabel("Time")
    plt.ylabel("Value")
    plt.grid(True)

!wget --no-check-certificate \
    https://raw.githubusercontent.com/jbrownlee/Datasets/master/daily-min-temperatures.csv \
    -O /tmp/daily-min-temperatures.csv

import csv
time_step = []
temps = []

with open('/tmp/daily-min-temperatures.csv') as csvfile:
    
    time_steps=

series = np.array(temps)
time = np.array(time_step)
plt.figure(figsize=(10, 6))
plot_series(time, series)

split_time = 2500
time_train = # YOUR CODE HERE
x_train = # YOUR CODE HERE
time_valid = # YOUR CODE HERE
x_valid = # YOUR CODE HERE

window_size = 30
batch_size = 32
shuffle_buffer_size = 1000

def windowed_dataset(series, window_size, batch_size, shuffle_buffer):
    # YOUR CODE HERE

def model_forecast(model, series, window_size):
    # YOUR CODE HERE

tf.keras.backend.clear_session()
tf.random.set_seed(51)
np.random.seed(51)
window_size = 64
batch_size = 256
train_set = windowed_dataset(x_train, window_size, batch_size, shuffle_buffer_size)
print(train_set)
print(x_train.shape)

model = tf.keras.models.Sequential([
# YOUR CODE HERE
])

lr_schedule = tf.keras.callbacks.LearningRateScheduler(
    lambda epoch: 1e-8 * 10**(epoch / 20))
optimizer = tf.keras.optimizers.SGD(lr=1e-8, momentum=0.9)
model.compile(loss=tf.keras.losses.Huber(),
              optimizer=optimizer,
              metrics=["mae"])
history = model.fit(train_set, epochs=100, callbacks=[lr_schedule])

WARNING: Logging before flag parsing goes to stderr.
W0719 05:10:05.389573 140234944071552 deprecation.py:323] From /usr/local/lib/python3.6/dist-packages/tensorflow/python/data/util/random_seed.py:58: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.where in 2.0, which has the same broadcast rule as np.where

<PrefetchDataset shapes: ((None, None, 1), (None, None, 1)), types: (tf.float64, tf.float64)>
(2500,)
Epoch 1/100
10/10 [==============================] - 6s 624ms/step - loss: 31.1549 - mae: 31.6551
Epoch 2/100
10/10 [==============================] - 4s 364ms/step - loss: 30.5696 - mae: 31.0771
Epoch 3/100
10/10 [==============================] - 4s 358ms/step - loss: 29.6691 - mae: 30.1811
Epoch 4/100
10/10 [==============================] - 4s 366ms/step - loss: 28.5431 - mae: 29.0596
Epoch 5/100
10/10 [==============================] - 4s 361ms/step - loss: 27.1744 - mae: 27.6976
Epoch 6/100
10/10 [==============================] - 4s 367ms/step - loss: 25.4676 - mae: 26.0015
Epoch 7/100
10/10 [==============================] - 4s 375ms/step - loss: 23.2987 - mae: 23.8487
Epoch 8/100
10/10 [==============================] - 4s 369ms/step - loss: 20.5506 - mae: 21.1192
Epoch 9/100
10/10 [==============================] - 4s 372ms/step - loss: 17.2408 - mae: 17.8223
Epoch 10/100
10/10 [==============================] - 4s 365ms/step - loss: 13.5549 - mae: 14.1350
Epoch 11/100
10/10 [==============================] - 4s 370ms/step - loss: 10.0753 - mae: 10.6367
Epoch 12/100
10/10 [==============================] - 4s 365ms/step - loss: 7.5638 - mae: 8.0985
Epoch 13/100
10/10 [==============================] - 4s 363ms/step - loss: 6.2445 - mae: 6.7619
Epoch 14/100
10/10 [==============================] - 4s 374ms/step - loss: 5.6749 - mae: 6.1854
Epoch 15/100
10/10 [==============================] - 4s 363ms/step - loss: 5.3090 - mae: 5.8168
Epoch 16/100
10/10 [==============================] - 4s 361ms/step - loss: 4.9122 - mae: 5.4169
Epoch 17/100
10/10 [==============================] - 4s 356ms/step - loss: 4.5318 - mae: 5.0305
Epoch 18/100
10/10 [==============================] - 4s 367ms/step - loss: 4.2115 - mae: 4.7068
Epoch 19/100
10/10 [==============================] - 4s 374ms/step - loss: 3.9429 - mae: 4.4360
Epoch 20/100
10/10 [==============================] - 4s 358ms/step - loss: 3.7309 - mae: 4.2194
Epoch 21/100
10/10 [==============================] - 4s 363ms/step - loss: 3.5706 - mae: 4.0551
Epoch 22/100
10/10 [==============================] - 4s 366ms/step - loss: 3.4527 - mae: 3.9344
Epoch 23/100
10/10 [==============================] - 4s 362ms/step - loss: 3.3617 - mae: 3.8423
Epoch 24/100
10/10 [==============================] - 4s 358ms/step - loss: 3.2876 - mae: 3.7666
Epoch 25/100
10/10 [==============================] - 3s 344ms/step - loss: 3.2224 - mae: 3.6997
Epoch 26/100
10/10 [==============================] - 4s 360ms/step - loss: 3.1596 - mae: 3.6359
Epoch 27/100
10/10 [==============================] - 4s 360ms/step - loss: 3.0964 - mae: 3.5717
Epoch 28/100
10/10 [==============================] - 4s 362ms/step - loss: 3.0322 - mae: 3.5064
Epoch 29/100
10/10 [==============================] - 4s 352ms/step - loss: 2.9662 - mae: 3.4392
Epoch 30/100
10/10 [==============================] - 4s 359ms/step - loss: 2.9004 - mae: 3.3720
Epoch 31/100
10/10 [==============================] - 4s 361ms/step - loss: 2.8376 - mae: 3.3081
Epoch 32/100
10/10 [==============================] - 4s 375ms/step - loss: 2.7775 - mae: 3.2475
Epoch 33/100
10/10 [==============================] - 4s 360ms/step - loss: 2.7202 - mae: 3.1899
Epoch 34/100
10/10 [==============================] - 4s 367ms/step - loss: 2.6662 - mae: 3.1360
Epoch 35/100
10/10 [==============================] - 4s 379ms/step - loss: 2.6152 - mae: 3.0848
Epoch 36/100
10/10 [==============================] - 4s 357ms/step - loss: 2.5663 - mae: 3.0353
Epoch 37/100
10/10 [==============================] - 4s 366ms/step - loss: 2.5192 - mae: 2.9872
Epoch 38/100
10/10 [==============================] - 4s 367ms/step - loss: 2.4735 - mae: 2.9408
Epoch 39/100
10/10 [==============================] - 4s 372ms/step - loss: 2.4296 - mae: 2.8964
Epoch 40/100
10/10 [==============================] - 4s 371ms/step - loss: 2.3873 - mae: 2.8534
Epoch 41/100
10/10 [==============================] - 4s 361ms/step - loss: 2.3463 - mae: 2.8119
Epoch 42/100
10/10 [==============================] - 4s 354ms/step - loss: 2.3060 - mae: 2.7707
Epoch 43/100
10/10 [==============================] - 4s 375ms/step - loss: 2.2663 - mae: 2.7301
Epoch 44/100
10/10 [==============================] - 4s 359ms/step - loss: 2.2269 - mae: 2.6899
Epoch 45/100
10/10 [==============================] - 4s 366ms/step - loss: 2.1898 - mae: 2.6519
Epoch 46/100
10/10 [==============================] - 4s 367ms/step - loss: 2.1563 - mae: 2.6181
Epoch 47/100
10/10 [==============================] - 4s 363ms/step - loss: 2.1248 - mae: 2.5861
Epoch 48/100
10/10 [==============================] - 4s 367ms/step - loss: 2.0958 - mae: 2.5568
Epoch 49/100
10/10 [==============================] - 4s 363ms/step - loss: 2.0688 - mae: 2.5296
Epoch 50/100
10/10 [==============================] - 4s 370ms/step - loss: 2.0442 - mae: 2.5045
Epoch 51/100
10/10 [==============================] - 4s 371ms/step - loss: 2.0220 - mae: 2.4818
Epoch 52/100
10/10 [==============================] - 4s 362ms/step - loss: 2.0018 - mae: 2.4611
Epoch 53/100
10/10 [==============================] - 4s 361ms/step - loss: 1.9801 - mae: 2.4393
Epoch 54/100
10/10 [==============================] - 4s 369ms/step - loss: 1.9586 - mae: 2.4171
Epoch 55/100
10/10 [==============================] - 4s 358ms/step - loss: 1.9390 - mae: 2.3972
Epoch 56/100
10/10 [==============================] - 4s 366ms/step - loss: 1.9186 - mae: 2.3763
Epoch 57/100
10/10 [==============================] - 4s 366ms/step - loss: 1.8975 - mae: 2.3550
Epoch 58/100
10/10 [==============================] - 4s 359ms/step - loss: 1.8743 - mae: 2.3320
Epoch 59/100
10/10 [==============================] - 4s 363ms/step - loss: 1.8738 - mae: 2.3310
Epoch 60/100
10/10 [==============================] - 4s 354ms/step - loss: 2.1527 - mae: 2.6201
Epoch 61/100
10/10 [==============================] - 4s 357ms/step - loss: 2.6764 - mae: 3.1247
Epoch 62/100
10/10 [==============================] - 4s 360ms/step - loss: 2.9935 - mae: 3.4806
Epoch 63/100
10/10 [==============================] - 4s 360ms/step - loss: 3.5219 - mae: 3.9875
Epoch 64/100
10/10 [==============================] - 4s 361ms/step - loss: 3.8284 - mae: 4.2965
Epoch 65/100
10/10 [==============================] - 4s 357ms/step - loss: 4.1265 - mae: 4.5856
Epoch 66/100
10/10 [==============================] - 4s 359ms/step - loss: 4.3062 - mae: 4.7664
Epoch 67/100
10/10 [==============================] - 4s 358ms/step - loss: 4.3039 - mae: 4.6838
Epoch 68/100
10/10 [==============================] - 4s 363ms/step - loss: 4.8341 - mae: 5.2257
Epoch 69/100
10/10 [==============================] - 4s 370ms/step - loss: 10.3352 - mae: 10.7649
Epoch 70/100
10/10 [==============================] - 4s 367ms/step - loss: 5.4020 - mae: 5.8328
Epoch 71/100
10/10 [==============================] - 4s 363ms/step - loss: 5.9798 - mae: 6.4835
Epoch 72/100
10/10 [==============================] - 4s 356ms/step - loss: 5.4958 - mae: 6.0146
Epoch 73/100
10/10 [==============================] - 4s 362ms/step - loss: 4.4955 - mae: 5.0201
Epoch 74/100
10/10 [==============================] - 4s 370ms/step - loss: 4.4764 - mae: 5.0171
Epoch 75/100
10/10 [==============================] - 4s 380ms/step - loss: 4.2825 - mae: 4.7287
Epoch 76/100
10/10 [==============================] - 4s 361ms/step - loss: 4.2044 - mae: 4.6470
Epoch 77/100
10/10 [==============================] - 4s 370ms/step - loss: 4.4160 - mae: 4.9125
Epoch 78/100
10/10 [==============================] - 4s 361ms/step - loss: 4.3770 - mae: 4.8831
Epoch 79/100
10/10 [==============================] - 4s 361ms/step - loss: 5.0487 - mae: 5.5839
Epoch 80/100
10/10 [==============================] - 4s 390ms/step - loss: 10.0358 - mae: 10.8604
Epoch 81/100
10/10 [==============================] - 4s 370ms/step - loss: 3.1176 - mae: 3.6060
Epoch 82/100
10/10 [==============================] - 4s 371ms/step - loss: 3.0097 - mae: 3.4891
Epoch 83/100
10/10 [==============================] - 4s 361ms/step - loss: 2.7912 - mae: 3.2609
Epoch 84/100
10/10 [==============================] - 4s 357ms/step - loss: 4.3135 - mae: 4.7803
Epoch 85/100
10/10 [==============================] - 4s 357ms/step - loss: 5.3703 - mae: 5.8508
Epoch 86/100
10/10 [==============================] - 4s 360ms/step - loss: 6.5221 - mae: 7.0175
Epoch 87/100
10/10 [==============================] - 4s 361ms/step - loss: 7.1154 - mae: 7.7249
Epoch 88/100
10/10 [==============================] - 4s 368ms/step - loss: 8.9975 - mae: 9.4580
Epoch 89/100
10/10 [==============================] - 4s 360ms/step - loss: 9.8069 - mae: 10.1397
Epoch 90/100
10/10 [==============================] - 4s 354ms/step - loss: 11.1364 - mae: 11.6797
Epoch 91/100
10/10 [==============================] - 4s 356ms/step - loss: 12.5922 - mae: 13.2602
Epoch 92/100
10/10 [==============================] - 4s 357ms/step - loss: 14.2512 - mae: 14.8289
Epoch 93/100
10/10 [==============================] - 4s 358ms/step - loss: 13.0192 - mae: 13.8809
Epoch 94/100
10/10 [==============================] - 4s 356ms/step - loss: 61.8923 - mae: 63.5737
Epoch 95/100
10/10 [==============================] - 4s 366ms/step - loss: 30.5821 - mae: 31.4688
Epoch 96/100
10/10 [==============================] - 4s 357ms/step - loss: 45.9889 - mae: 46.9740
Epoch 97/100
10/10 [==============================] - 4s 355ms/step - loss: 51.7050 - mae: 51.8100
Epoch 98/100
10/10 [==============================] - 4s 361ms/step - loss: 59.1678 - mae: 57.9463
Epoch 99/100
10/10 [==============================] - 4s 373ms/step - loss: 66.1686 - mae: 68.4903
Epoch 100/100
10/10 [==============================] - 4s 360ms/step - loss: 74.1415 - mae: 72.0624

plt.semilogx(history.history["lr"], history.history["loss"])
plt.axis([1e-8, 1e-4, 0, 60])

[1e-08, 0.0001, 0, 60]

tf.keras.backend.clear_session()
tf.random.set_seed(51)
np.random.seed(51)
train_set = windowed_dataset(x_train, window_size=60, batch_size=100, shuffle_buffer=shuffle_buffer_size)
model = tf.keras.models.Sequential([
# YOUR CODE HERE
])


optimizer = tf.keras.optimizers.SGD(lr=# YOUR CODE HERE, momentum=0.9)
model.compile(loss=tf.keras.losses.Huber(),
              optimizer=optimizer,
              metrics=["mae"])
history = model.fit(train_set,epochs=# YOUR CODE HERE)
                                    
# EXPECTED OUTPUT SHOULD SEE AN MAE OF <2 WITHIN ABOUT 30 EPOCHS

Epoch 1/150
25/25 [==============================] - 6s 243ms/step - loss: 9.9624 - mae: 10.5789
Epoch 2/150
25/25 [==============================] - 3s 136ms/step - loss: 2.5390 - mae: 3.0130
Epoch 3/150
25/25 [==============================] - 3s 131ms/step - loss: 1.9265 - mae: 2.3815
Epoch 4/150
25/25 [==============================] - 3s 137ms/step - loss: 1.8597 - mae: 2.3125
Epoch 5/150
25/25 [==============================] - 3s 139ms/step - loss: 1.8181 - mae: 2.2696
Epoch 6/150
25/25 [==============================] - 3s 140ms/step - loss: 1.7882 - mae: 2.2385
Epoch 7/150
25/25 [==============================] - 4s 141ms/step - loss: 1.7618 - mae: 2.2112
Epoch 8/150
25/25 [==============================] - 3s 135ms/step - loss: 1.7382 - mae: 2.1870
Epoch 9/150
25/25 [==============================] - 3s 136ms/step - loss: 1.7167 - mae: 2.1650
Epoch 10/150
25/25 [==============================] - 3s 136ms/step - loss: 1.6976 - mae: 2.1454
Epoch 11/150
25/25 [==============================] - 3s 138ms/step - loss: 1.6808 - mae: 2.1281
Epoch 12/150
25/25 [==============================] - 3s 138ms/step - loss: 1.6661 - mae: 2.1128
Epoch 13/150
25/25 [==============================] - 4s 142ms/step - loss: 1.6531 - mae: 2.0993
Epoch 14/150
25/25 [==============================] - 4s 142ms/step - loss: 1.6417 - mae: 2.0872
Epoch 15/150
25/25 [==============================] - 3s 135ms/step - loss: 1.6315 - mae: 2.0764
Epoch 16/150
25/25 [==============================] - 3s 128ms/step - loss: 1.6223 - mae: 2.0667
Epoch 17/150
25/25 [==============================] - 3s 132ms/step - loss: 1.6141 - mae: 2.0579
Epoch 18/150
25/25 [==============================] - 3s 130ms/step - loss: 1.6067 - mae: 2.0500
Epoch 19/150
25/25 [==============================] - 3s 133ms/step - loss: 1.6000 - mae: 2.0429
Epoch 20/150
25/25 [==============================] - 3s 138ms/step - loss: 1.5939 - mae: 2.0364
Epoch 21/150
25/25 [==============================] - 3s 137ms/step - loss: 1.5883 - mae: 2.0306
Epoch 22/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5833 - mae: 2.0254
Epoch 23/150
25/25 [==============================] - 4s 141ms/step - loss: 1.5787 - mae: 2.0207
Epoch 24/150
25/25 [==============================] - 3s 140ms/step - loss: 1.5745 - mae: 2.0163
Epoch 25/150
25/25 [==============================] - 4s 141ms/step - loss: 1.5707 - mae: 2.0124
Epoch 26/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5672 - mae: 2.0089
Epoch 27/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5640 - mae: 2.0056
Epoch 28/150
25/25 [==============================] - 3s 135ms/step - loss: 1.5610 - mae: 2.0026
Epoch 29/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5583 - mae: 1.9998
Epoch 30/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5558 - mae: 1.9973
Epoch 31/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5534 - mae: 1.9949
Epoch 32/150
25/25 [==============================] - 3s 139ms/step - loss: 1.5512 - mae: 1.9927
Epoch 33/150
25/25 [==============================] - 3s 139ms/step - loss: 1.5491 - mae: 1.9906
Epoch 34/150
25/25 [==============================] - 3s 138ms/step - loss: 1.5472 - mae: 1.9886
Epoch 35/150
25/25 [==============================] - 4s 140ms/step - loss: 1.5453 - mae: 1.9868
Epoch 36/150
25/25 [==============================] - 4s 141ms/step - loss: 1.5436 - mae: 1.9850
Epoch 37/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5419 - mae: 1.9833
Epoch 38/150
25/25 [==============================] - 3s 130ms/step - loss: 1.5404 - mae: 1.9817
Epoch 39/150
25/25 [==============================] - 3s 130ms/step - loss: 1.5389 - mae: 1.9803
Epoch 40/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5376 - mae: 1.9789
Epoch 41/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5363 - mae: 1.9776
Epoch 42/150
25/25 [==============================] - 3s 129ms/step - loss: 1.5350 - mae: 1.9764
Epoch 43/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5338 - mae: 1.9752
Epoch 44/150
25/25 [==============================] - 4s 140ms/step - loss: 1.5327 - mae: 1.9741
Epoch 45/150
25/25 [==============================] - 4s 141ms/step - loss: 1.5316 - mae: 1.9730
Epoch 46/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5305 - mae: 1.9719
Epoch 47/150
25/25 [==============================] - 3s 135ms/step - loss: 1.5295 - mae: 1.9708
Epoch 48/150
25/25 [==============================] - 3s 138ms/step - loss: 1.5286 - mae: 1.9698
Epoch 49/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5276 - mae: 1.9689
Epoch 50/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5267 - mae: 1.9680
Epoch 51/150
25/25 [==============================] - 3s 139ms/step - loss: 1.5259 - mae: 1.9672
Epoch 52/150
25/25 [==============================] - 3s 134ms/step - loss: 1.5251 - mae: 1.9664
Epoch 53/150
25/25 [==============================] - 3s 135ms/step - loss: 1.5243 - mae: 1.9656
Epoch 54/150
25/25 [==============================] - 3s 137ms/step - loss: 1.5235 - mae: 1.9648
Epoch 55/150
25/25 [==============================] - 3s 137ms/step - loss: 1.5228 - mae: 1.9640
Epoch 56/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5221 - mae: 1.9633
Epoch 57/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5213 - mae: 1.9626
Epoch 58/150
25/25 [==============================] - 3s 137ms/step - loss: 1.5206 - mae: 1.9619
Epoch 59/150
25/25 [==============================] - 3s 135ms/step - loss: 1.5199 - mae: 1.9611
Epoch 60/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5193 - mae: 1.9604
Epoch 61/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5186 - mae: 1.9597
Epoch 62/150
25/25 [==============================] - 3s 131ms/step - loss: 1.5179 - mae: 1.9589
Epoch 63/150
25/25 [==============================] - 3s 131ms/step - loss: 1.5173 - mae: 1.9582
Epoch 64/150
25/25 [==============================] - 3s 135ms/step - loss: 1.5166 - mae: 1.9576
Epoch 65/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5161 - mae: 1.9570
Epoch 66/150
25/25 [==============================] - 3s 138ms/step - loss: 1.5155 - mae: 1.9564
Epoch 67/150
25/25 [==============================] - 3s 140ms/step - loss: 1.5149 - mae: 1.9557
Epoch 68/150
25/25 [==============================] - 4s 141ms/step - loss: 1.5142 - mae: 1.9550
Epoch 69/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5136 - mae: 1.9543
Epoch 70/150
25/25 [==============================] - 3s 130ms/step - loss: 1.5129 - mae: 1.9536
Epoch 71/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5122 - mae: 1.9528
Epoch 72/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5115 - mae: 1.9521
Epoch 73/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5108 - mae: 1.9514
Epoch 74/150
25/25 [==============================] - 3s 135ms/step - loss: 1.5102 - mae: 1.9508
Epoch 75/150
25/25 [==============================] - 3s 133ms/step - loss: 1.5097 - mae: 1.9503
Epoch 76/150
25/25 [==============================] - 3s 134ms/step - loss: 1.5091 - mae: 1.9497
Epoch 77/150
25/25 [==============================] - 3s 137ms/step - loss: 1.5084 - mae: 1.9491
Epoch 78/150
25/25 [==============================] - 3s 128ms/step - loss: 1.5078 - mae: 1.9485
Epoch 79/150
25/25 [==============================] - 3s 129ms/step - loss: 1.5072 - mae: 1.9478
Epoch 80/150
25/25 [==============================] - 3s 134ms/step - loss: 1.5065 - mae: 1.9471
Epoch 81/150
25/25 [==============================] - 3s 136ms/step - loss: 1.5059 - mae: 1.9465
Epoch 82/150
25/25 [==============================] - 3s 134ms/step - loss: 1.5053 - mae: 1.9459
Epoch 83/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5048 - mae: 1.9454
Epoch 84/150
25/25 [==============================] - 3s 134ms/step - loss: 1.5043 - mae: 1.9448
Epoch 85/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5038 - mae: 1.9443
Epoch 86/150
25/25 [==============================] - 3s 131ms/step - loss: 1.5033 - mae: 1.9438
Epoch 87/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5028 - mae: 1.9433
Epoch 88/150
25/25 [==============================] - 3s 130ms/step - loss: 1.5023 - mae: 1.9428
Epoch 89/150
25/25 [==============================] - 3s 134ms/step - loss: 1.5017 - mae: 1.9422
Epoch 90/150
25/25 [==============================] - 3s 132ms/step - loss: 1.5012 - mae: 1.9416
Epoch 91/150
25/25 [==============================] - 3s 131ms/step - loss: 1.5004 - mae: 1.9408
Epoch 92/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4992 - mae: 1.9396
Epoch 93/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4982 - mae: 1.9386
Epoch 94/150
25/25 [==============================] - 3s 136ms/step - loss: 1.4975 - mae: 1.9378
Epoch 95/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4968 - mae: 1.9371
Epoch 96/150
25/25 [==============================] - 3s 136ms/step - loss: 1.4962 - mae: 1.9364
Epoch 97/150
25/25 [==============================] - 3s 137ms/step - loss: 1.4955 - mae: 1.9358
Epoch 98/150
25/25 [==============================] - 3s 140ms/step - loss: 1.4949 - mae: 1.9351
Epoch 99/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4942 - mae: 1.9344
Epoch 100/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4934 - mae: 1.9336
Epoch 101/150
25/25 [==============================] - 3s 135ms/step - loss: 1.4927 - mae: 1.9328
Epoch 102/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4919 - mae: 1.9319
Epoch 103/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4913 - mae: 1.9313
Epoch 104/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4908 - mae: 1.9307
Epoch 105/150
25/25 [==============================] - 3s 133ms/step - loss: 1.4903 - mae: 1.9302
Epoch 106/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4899 - mae: 1.9298
Epoch 107/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4895 - mae: 1.9293
Epoch 108/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4892 - mae: 1.9289
Epoch 109/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4888 - mae: 1.9285
Epoch 110/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4884 - mae: 1.9282
Epoch 111/150
25/25 [==============================] - 3s 133ms/step - loss: 1.4881 - mae: 1.9278
Epoch 112/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4878 - mae: 1.9275
Epoch 113/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4874 - mae: 1.9271
Epoch 114/150
25/25 [==============================] - 3s 136ms/step - loss: 1.4871 - mae: 1.9268
Epoch 115/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4868 - mae: 1.9265
Epoch 116/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4865 - mae: 1.9262
Epoch 117/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4862 - mae: 1.9258
Epoch 118/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4859 - mae: 1.9255
Epoch 119/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4856 - mae: 1.9252
Epoch 120/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4853 - mae: 1.9249
Epoch 121/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4851 - mae: 1.9247
Epoch 122/150
25/25 [==============================] - 3s 135ms/step - loss: 1.4848 - mae: 1.9243
Epoch 123/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4845 - mae: 1.9240
Epoch 124/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4842 - mae: 1.9238
Epoch 125/150
25/25 [==============================] - 3s 133ms/step - loss: 1.4840 - mae: 1.9235
Epoch 126/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4838 - mae: 1.9232
Epoch 127/150
25/25 [==============================] - 3s 129ms/step - loss: 1.4835 - mae: 1.9230
Epoch 128/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4833 - mae: 1.9228
Epoch 129/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4831 - mae: 1.9225
Epoch 130/150
25/25 [==============================] - 3s 129ms/step - loss: 1.4828 - mae: 1.9222
Epoch 131/150
25/25 [==============================] - 3s 128ms/step - loss: 1.4826 - mae: 1.9220
Epoch 132/150
25/25 [==============================] - 3s 127ms/step - loss: 1.4823 - mae: 1.9217
Epoch 133/150
25/25 [==============================] - 3s 129ms/step - loss: 1.4821 - mae: 1.9214
Epoch 134/150
25/25 [==============================] - 3s 129ms/step - loss: 1.4818 - mae: 1.9212
Epoch 135/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4815 - mae: 1.9209
Epoch 136/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4813 - mae: 1.9206
Epoch 137/150
25/25 [==============================] - 3s 127ms/step - loss: 1.4811 - mae: 1.9204
Epoch 138/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4808 - mae: 1.9201
Epoch 139/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4806 - mae: 1.9199
Epoch 140/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4804 - mae: 1.9197
Epoch 141/150
25/25 [==============================] - 3s 132ms/step - loss: 1.4802 - mae: 1.9194
Epoch 142/150
25/25 [==============================] - 3s 129ms/step - loss: 1.4799 - mae: 1.9192
Epoch 143/150
25/25 [==============================] - 3s 134ms/step - loss: 1.4797 - mae: 1.9189
Epoch 144/150
25/25 [==============================] - 3s 135ms/step - loss: 1.4795 - mae: 1.9187
Epoch 145/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4794 - mae: 1.9185
Epoch 146/150
25/25 [==============================] - 3s 131ms/step - loss: 1.4791 - mae: 1.9183
Epoch 147/150
25/25 [==============================] - 3s 130ms/step - loss: 1.4789 - mae: 1.9180
Epoch 148/150
25/25 [==============================] - 3s 136ms/step - loss: 1.4787 - mae: 1.9178
Epoch 149/150
25/25 [==============================] - 3s 137ms/step - loss: 1.4784 - mae: 1.9175
Epoch 150/150
25/25 [==============================] - 4s 143ms/step - loss: 1.4782 - mae: 1.9173

rnn_forecast = model_forecast(model, series[..., np.newaxis], window_size)
rnn_forecast = rnn_forecast[split_time - window_size:-1, -1, 0]

plt.figure(figsize=(10, 6))
plot_series(time_valid, x_valid)
plot_series(time_valid, rnn_forecast)

# EXPECTED OUTPUT. PLOT SHOULD SHOW PROJECTIONS FOLLOWING ORIGINAL DATA CLOSELY

tf.keras.metrics.mean_absolute_error(x_valid, rnn_forecast).numpy()

# EXPECTED OUTPUT MAE < 2 -- I GOT 1.789626

1.780626

print(rnn_forecast)
# EXPECTED OUTPUT -- ARRAY OF VALUES IN THE LOW TEENS

[11.636601 10.97607  12.159701 ... 13.589686 13.726407 14.940471]