In [0]:

    

#@title Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


    

In [0]:

    

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

keras = tf.keras


    

In [0]:

    

def plot_series(time, series, format="-", start=0, end=None, label=None):
    plt.plot(time[start:end], series[start:end], format, label=label)
    plt.xlabel("Time")
    plt.ylabel("Value")
    if label:
        plt.legend(fontsize=14)
    plt.grid(True)
    
def trend(time, slope=0):
    return slope * time
  
  
def seasonal_pattern(season_time):
    """Just an arbitrary pattern, you can change it if you wish"""
    return np.where(season_time < 0.4,
                    np.cos(season_time * 2 * np.pi),
                    1 / np.exp(3 * season_time))

  
def seasonality(time, period, amplitude=1, phase=0):
    """Repeats the same pattern at each period"""
    season_time = ((time + phase) % period) / period
    return amplitude * seasonal_pattern(season_time)
  
  
def white_noise(time, noise_level=1, seed=None):
    rnd = np.random.RandomState(seed)
    return rnd.randn(len(time)) * noise_level


    

In [0]:

    

time = np.arange(4 * 365 + 1)

slope = 0.05
baseline = 10
amplitude = 40
series = baseline + trend(time, slope) + seasonality(time, period=365, amplitude=amplitude)

noise_level = 5
noise = white_noise(time, noise_level, seed=42)

series += noise

plt.figure(figsize=(10, 6))
plot_series(time, series)
plt.show()


    

In [0]:

    

split_time = 1000
time_train = time[:split_time]
x_train = series[:split_time]
time_valid = time[split_time:]
x_valid = series[split_time:]


    

In [0]:

    

def sequential_window_dataset(series, window_size):
    series = tf.expand_dims(series, axis=-1)
    ds = tf.data.Dataset.from_tensor_slices(series)
    ds = ds.window(window_size + 1, shift=window_size, drop_remainder=True)
    ds = ds.flat_map(lambda window: window.batch(window_size + 1))
    ds = ds.map(lambda window: (window[:-1], window[1:]))
    return ds.batch(1).prefetch(1)


    

In [0]:

    

for X_batch, y_batch in sequential_window_dataset(tf.range(10), 3):
    print(X_batch.numpy(), y_batch.numpy())


    

In [0]:

    

class ResetStatesCallback(keras.callbacks.Callback):
    def on_epoch_begin(self, epoch, logs):
        self.model.reset_states()


    

In [0]:

    

keras.backend.clear_session()
tf.random.set_seed(42)
np.random.seed(42)

window_size = 30
train_set = sequential_window_dataset(x_train, window_size)

model = keras.models.Sequential([
  keras.layers.SimpleRNN(100, return_sequences=True, stateful=True,
                         batch_input_shape=[1, None, 1]),
  keras.layers.SimpleRNN(100, return_sequences=True, stateful=True),
  keras.layers.Dense(1),
  keras.layers.Lambda(lambda x: x * 200.0)
])
lr_schedule = keras.callbacks.LearningRateScheduler(
    lambda epoch: 1e-8 * 10**(epoch / 30))
reset_states = ResetStatesCallback()
optimizer = keras.optimizers.SGD(lr=1e-8, momentum=0.9)
model.compile(loss=keras.losses.Huber(),
              optimizer=optimizer,
              metrics=["mae"])
history = model.fit(train_set, epochs=100,
                    callbacks=[lr_schedule, reset_states])


    

In [0]:

    

plt.semilogx(history.history["lr"], history.history["loss"])
plt.axis([1e-8, 1e-4, 0, 30])


    

In [0]:

    

keras.backend.clear_session()
tf.random.set_seed(42)
np.random.seed(42)

window_size = 30
train_set = sequential_window_dataset(x_train, window_size)
valid_set = sequential_window_dataset(x_valid, window_size)

model = keras.models.Sequential([
  keras.layers.SimpleRNN(100, return_sequences=True, stateful=True,
                         batch_input_shape=[1, None, 1]),
  keras.layers.SimpleRNN(100, return_sequences=True, stateful=True),
  keras.layers.Dense(1),
  keras.layers.Lambda(lambda x: x * 200.0)
])
optimizer = keras.optimizers.SGD(lr=1e-7, momentum=0.9)
model.compile(loss=keras.losses.Huber(),
              optimizer=optimizer,
              metrics=["mae"])
reset_states = ResetStatesCallback()
model_checkpoint = keras.callbacks.ModelCheckpoint(
    "my_checkpoint.h5", save_best_only=True)
early_stopping = keras.callbacks.EarlyStopping(patience=50)
model.fit(train_set, epochs=500,
          validation_data=valid_set,
          callbacks=[early_stopping, model_checkpoint, reset_states])


    

In [0]:

    

model = keras.models.load_model("my_checkpoint.h5")


    

In [0]:

    

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


    

In [0]:

    

rnn_forecast.shape


    

In [0]:

    

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


    

In [0]:

    

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