Predicción de secuencia de valores futuros de una acción. Apple (LSTM)#

Importar las librería requeridas#

#
from __future__ import absolute_import, division, print_function, unicode_literals

import tensorflow as tf

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
#%matplotlib inline
from sklearn.preprocessing import MinMaxScaler

#setting figure size
from matplotlib.pylab import rcParams
rcParams['figure.figsize'] = 20,10

#importing required libraries
from sklearn.preprocessing import MinMaxScaler

# importa objetos de keras
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Input, Dense, Dropout, LSTM
print("Versión de Tensorflow: ", tf.__version__)

# optimizador
from tensorflow.keras.optimizers import Adam

Versión de Tensorflow:  2.9.1

Lectura de los datos#

Estos datos corresponden a la empresa Apple. Son 3019 datos que corresponden a observaciones del precio de la acción, el número de transacciones de la acción (compra-venta). Los datos son diarios (dias hábiles o comerciales). Están entre el 3 de enero de 2006 hasta el 1 de enero de 2018.

la columna Date es la fecha, Open es el valor de acción a la apertura del mercado, High el valor más alto alcanzado en el día, Low el valor más bajo del día, Close el valor al cierre, Volume es el volúmenes de acciones transadas en el día y Name es el código de identificación de la empresa, Apple en este caso.

Los datos puede ser bajados directamente de Kaggle

#reading from a local file

df = pd.read_csv('https://raw.githubusercontent.com/AprendizajeProfundo/Libro-Fundamentos/main/Redes_Recurrentes/Datos/AAPL_2006-01-01_to_2018-01-01.csv')

# looking at the first five rows of the data
print('\n Shape of the data:')
print(df.shape)
df.head()

 Shape of the data:
(3019, 7)
Date Open High Low Close Volume Name
0 2006-01-03 10.34 10.68 10.32 10.68 201853036 AAPL
1 2006-01-04 10.73 10.85 10.64 10.71 155225609 AAPL
2 2006-01-05 10.69 10.70 10.54 10.63 112396081 AAPL
3 2006-01-06 10.75 10.96 10.65 10.90 176139334 AAPL
4 2006-01-09 10.96 11.03 10.82 10.86 168861224 AAPL

Vamos a cambiar el índice de los datos. Tomaremos la fecha como indice: df.index. Los datos se reordenan para invertir la tabla, debido a que los datos contienen las observaciones más recientes en la parte superior de la tabla.

Extrae datos para la serie que se desea predecir-close#

#creating dataframe with date and the target variable

df['Date'] = pd.to_datetime(df.Date,format='%Y-%m-%d')
df.index = df['Date']
# df = df.sort_index(ascending=True, axis=0)
data = pd.DataFrame(df[['Date', 'Close']])
#
#setting index
data.index = data.Date
data.drop('Date', axis=1, inplace=True)
data.head()
Close
Date
2006-01-03 10.68
2006-01-04 10.71
2006-01-05 10.63
2006-01-06 10.90
2006-01-09 10.86

Visualización de la serie precio al cierre#

# plot
len_data = len(data)
len_train = int(len_data*0.8) # 80%  = 3019
len_test = len_data- len_train # 20% = 2415
print (len_data, '=', len_train, '+',len_test)

3019 = 2415 + 604

plt.figure(figsize=(16,8))
plt.plot(data[:len_train], label='Conjunto de entrenamiento (Training set): {} puntos (80%)'.format(len_train))
plt.plot(data['Close'][len_train:], label='Conjunto de validación (Validation set): {} puntos (20%)'.format(len_test)) #248 data
plt.title("Apple: Historia del precio la acción al cierre (Close)", size = 20)
plt.legend()
plt.show()
../../_images/rnr_accion_Apple_Prediccion_tres_dias-multiple_12_0.png

Preparación de los datos para el entrenamiento de la red LSTM#

Para evitar problemas con las tendencias y para mejorar la estimación (entrenamiento) los datos se van a transformar a la escala \([0,1]\). Para las predicciones se utiliza la transformación inversa.

Primero extrae los valores y se crea el objeto MinMaxScaler#

#creating train and test sets
dataset = data.values


# create the scaler object and scale the data
scaler = MinMaxScaler(feature_range=(0, 1))
#scaled_data = np.array(scaler.fit_transform(dataset))
dataset = np.squeeze(np.array(scaler.fit_transform(dataset)),axis=1)
# dataset = pd.DataFrame(scaled_data,index=data.index, columns=['serie'])

dataset.shape

(3019,)

Crea datos de entrenamiento#

La red LSTM tendrá como entrada «time_step» datos consecutivos, y como salida 5 datos (la predicción a partir de esos «time_step» datos se hace para los siguentes 5 días). Se conformará de esta forma el set de entrenamiento

  1. Número de datos consecutivos para entrenamiento: time_step = 60.

  2. Días a predecir: days = 1

Función para crear los datos de entrenamiento#

def multipaso_data(dataset, target, start_index, end_index, history_size,
                      target_size,  single_step=False):
    ''' dataset: conjunto de datos para las secuencias de entrada
        target:  conjunto de datos para las secuencias de salida
        start_index: índice inicial de donde empezar a tomar los datos
        end_index: índice final para tomar los datos. None para tomarlos todos
        history_size: tamaño de la venytana para crear las secuencias
        target_size: dentro de cuántas observaciones futuras desea pronosticar
        single_step: Predecir solamente un valor futuro (=True),
                     o predecir todos los valores hasta target_size(=False)
    '''  
    data = []
    labels = []

    start_index = start_index + history_size
    if end_index is None:
        end_index = len(dataset) - target_size

    for i in range(start_index, end_index):
        indices = range(i-history_size, i)
        data.append(dataset[indices])

        if single_step:
            labels.append(target[i+target_size])
        else:
            labels.append(target[i:i+target_size])

    return np.array(data), np.array(labels)

Se coloca una semilla para gerantizar reproductibidad

tf.random.set_seed(100)

#
# hiperparámetros para crear las secuencias
#
# tamaño de pasos a futuro
future_target = 10

# tamaño secuencias de entrada
past_history = 60 

TRAIN_SPLIT = int(len_data*0.8) #2415: nuḿer0 de datos entreno

# Crea los datos
X_train, y_train = multipaso_data(dataset, dataset, 0,
                                                 TRAIN_SPLIT, past_history,
                                                 future_target)
X_test, y_test = multipaso_data(dataset, dataset, TRAIN_SPLIT,
                                                 None, past_history,
                                                 future_target)

print(TRAIN_SPLIT)
print(X_train.shape)
print(y_train.shape)
print(X_test.shape)
print(y_test.shape)

2415
(2355, 60)
(2355, 10)
(534, 60)
(534, 10)

Crea el modelo LSTM#

Omitimos esta sección. Usaremso el modelo entrenado a un día

# shapes
input_shape = (X_train.shape[1], 1)
units = 50

# layers
inputs = Input(input_shape)
x = Dropout(0.0, name= 'Dropout_01')(inputs)
x = LSTM(units=units, name='LSTM_layer')(x)
#x = LSTM(units=units, return_sequences=True,name='LSTM_layer')(inputs)
#x = Dropout(0.4)
#x = LSTM(units=units//2, name='LSTM_layer_2')(x)
#x = Dropout(0.4)
x = Dropout(0.0, name= 'Dropout_02')(x)
outputs = Dense(future_target)(x)

# model
model_01 = Model(inputs=inputs, outputs=outputs, name='series_LSTM_model')
model_01.summary()

Model: "series_LSTM_model"

_________________________________________________________________

 Layer (type)                Output Shape              Param #   

=================================================================

 input_1 (InputLayer)        [(None, 60, 1)]           0         

                                                                 

 Dropout_01 (Dropout)        (None, 60, 1)             0         

                                                                 

 LSTM_layer (LSTM)           (None, 50)                10400     

                                                                 

 Dropout_02 (Dropout)        (None, 50)                0         

                                                                 

 dense (Dense)               (None, 10)                510       

                                                                 

=================================================================

Total params: 10,910

Trainable params: 10,910

Non-trainable params: 0

_________________________________________________________________

Compila#

Se usará el optimizador Adam y la función de pérdida MSE

model_01.compile(loss='mean_squared_error',
  optimizer=Adam(0.001))

Lee modelo preentrenado a un día#

Entrena el modelo#

#history = model_01.fit(X_train,y_train,epochs=20,batch_size=32)
history = model_01.fit(
    X_train, y_train,
    epochs=40,
    batch_size=32,
    validation_split=0.1,
    verbose=1,
    shuffle=False
)

Epoch 1/40

 1/67 [..............................] - ETA: 3:55 - loss: 1.9991e-04


 3/67 [>.............................] - ETA: 1s - loss: 8.8490e-05  


 5/67 [=>............................] - ETA: 1s - loss: 1.1145e-04


 7/67 [==>...........................] - ETA: 1s - loss: 1.4138e-04


 9/67 [===>..........................] - ETA: 1s - loss: 1.6737e-04


12/67 [====>.........................] - ETA: 1s - loss: 3.8926e-04


14/67 [=====>........................] - ETA: 1s - loss: 6.7570e-04


16/67 [======>.......................] - ETA: 1s - loss: 6.6382e-04


18/67 [=======>......................] - ETA: 1s - loss: 6.6036e-04


20/67 [=======>......................] - ETA: 1s - loss: 7.1802e-04


22/67 [========>.....................] - ETA: 1s - loss: 8.9220e-04


25/67 [==========>...................] - ETA: 1s - loss: 8.8173e-04


27/67 [===========>..................] - ETA: 1s - loss: 8.8006e-04


29/67 [===========>..................] - ETA: 1s - loss: 0.0010    


31/67 [============>.................] - ETA: 1s - loss: 0.0012


33/67 [=============>................] - ETA: 0s - loss: 0.0016


35/67 [==============>...............] - ETA: 0s - loss: 0.0018


37/67 [===============>..............] - ETA: 0s - loss: 0.0021


40/67 [================>.............] - ETA: 0s - loss: 0.0022


43/67 [==================>...........] - ETA: 0s - loss: 0.0022


46/67 [===================>..........] - ETA: 0s - loss: 0.0021


49/67 [====================>.........] - ETA: 0s - loss: 0.0025


52/67 [======================>.......] - ETA: 0s - loss: 0.0025


55/67 [=======================>......] - ETA: 0s - loss: 0.0030


57/67 [========================>.....] - ETA: 0s - loss: 0.0030


59/67 [=========================>....] - ETA: 0s - loss: 0.0030


61/67 [==========================>...] - ETA: 0s - loss: 0.0033


63/67 [===========================>..] - ETA: 0s - loss: 0.0035


65/67 [============================>.] - ETA: 0s - loss: 0.0040


67/67 [==============================] - ETA: 0s - loss: 0.0044


67/67 [==============================] - 6s 43ms/step - loss: 0.0044 - val_loss: 0.0273

Epoch 2/40

 1/67 [..............................] - ETA: 2s - loss: 0.0328


 3/67 [>.............................] - ETA: 1s - loss: 0.0366


 5/67 [=>............................] - ETA: 2s - loss: 0.0356


 7/67 [==>...........................] - ETA: 2s - loss: 0.0328


 9/67 [===>..........................] - ETA: 2s - loss: 0.0292


11/67 [===>..........................] - ETA: 1s - loss: 0.0253


13/67 [====>.........................] - ETA: 1s - loss: 0.0218


15/67 [=====>........................] - ETA: 1s - loss: 0.0192


17/67 [======>.......................] - ETA: 1s - loss: 0.0170


19/67 [=======>......................] - ETA: 1s - loss: 0.0152


21/67 [========>.....................] - ETA: 1s - loss: 0.0139


23/67 [=========>....................] - ETA: 1s - loss: 0.0127


25/67 [==========>...................] - ETA: 1s - loss: 0.0117


27/67 [===========>..................] - ETA: 1s - loss: 0.0109


30/67 [============>.................] - ETA: 1s - loss: 0.0101


33/67 [=============>................] - ETA: 0s - loss: 0.0097


36/67 [===============>..............] - ETA: 0s - loss: 0.0094


38/67 [================>.............] - ETA: 0s - loss: 0.0094


40/67 [================>.............] - ETA: 0s - loss: 0.0093


42/67 [=================>............] - ETA: 0s - loss: 0.0092


44/67 [==================>...........] - ETA: 0s - loss: 0.0091


46/67 [===================>..........] - ETA: 0s - loss: 0.0089


48/67 [====================>.........] - ETA: 0s - loss: 0.0091


50/67 [=====================>........] - ETA: 0s - loss: 0.0090


52/67 [======================>.......] - ETA: 0s - loss: 0.0088


54/67 [=======================>......] - ETA: 0s - loss: 0.0087


56/67 [========================>.....] - ETA: 0s - loss: 0.0087


58/67 [========================>.....] - ETA: 0s - loss: 0.0085


60/67 [=========================>....] - ETA: 0s - loss: 0.0083


62/67 [==========================>...] - ETA: 0s - loss: 0.0081


64/67 [===========================>..] - ETA: 0s - loss: 0.0079


66/67 [============================>.] - ETA: 0s - loss: 0.0079


67/67 [==============================] - 2s 30ms/step - loss: 0.0079 - val_loss: 0.0122

Epoch 3/40

 1/67 [..............................] - ETA: 1s - loss: 0.0139


 3/67 [>.............................] - ETA: 1s - loss: 0.0149


 5/67 [=>............................] - ETA: 1s - loss: 0.0145


 7/67 [==>...........................] - ETA: 1s - loss: 0.0139


10/67 [===>..........................] - ETA: 1s - loss: 0.0122


13/67 [====>.........................] - ETA: 1s - loss: 0.0103


15/67 [=====>........................] - ETA: 1s - loss: 0.0092


18/67 [=======>......................] - ETA: 1s - loss: 0.0078


20/67 [=======>......................] - ETA: 1s - loss: 0.0071


22/67 [========>.....................] - ETA: 1s - loss: 0.0066


24/67 [=========>....................] - ETA: 1s - loss: 0.0061


26/67 [==========>...................] - ETA: 1s - loss: 0.0057


28/67 [===========>..................] - ETA: 1s - loss: 0.0053


31/67 [============>.................] - ETA: 0s - loss: 0.0049


33/67 [=============>................] - ETA: 0s - loss: 0.0047


35/67 [==============>...............] - ETA: 0s - loss: 0.0046


38/67 [================>.............] - ETA: 0s - loss: 0.0046


40/67 [================>.............] - ETA: 0s - loss: 0.0045


42/67 [=================>............] - ETA: 0s - loss: 0.0045


44/67 [==================>...........] - ETA: 0s - loss: 0.0044


47/67 [====================>.........] - ETA: 0s - loss: 0.0043


50/67 [=====================>........] - ETA: 0s - loss: 0.0042


53/67 [======================>.......] - ETA: 0s - loss: 0.0042


55/67 [=======================>......] - ETA: 0s - loss: 0.0041


57/67 [========================>.....] - ETA: 0s - loss: 0.0041


59/67 [=========================>....] - ETA: 0s - loss: 0.0039


62/67 [==========================>...] - ETA: 0s - loss: 0.0038


65/67 [============================>.] - ETA: 0s - loss: 0.0038


67/67 [==============================] - ETA: 0s - loss: 0.0038


67/67 [==============================] - 2s 29ms/step - loss: 0.0038 - val_loss: 0.0037

Epoch 4/40

 1/67 [..............................] - ETA: 2s - loss: 0.0052


 4/67 [>.............................] - ETA: 1s - loss: 0.0056


 7/67 [==>...........................] - ETA: 1s - loss: 0.0053


 9/67 [===>..........................] - ETA: 1s - loss: 0.0049


12/67 [====>.........................] - ETA: 1s - loss: 0.0043


14/67 [=====>........................] - ETA: 1s - loss: 0.0039


16/67 [======>.......................] - ETA: 1s - loss: 0.0036


18/67 [=======>......................] - ETA: 1s - loss: 0.0032


20/67 [=======>......................] - ETA: 1s - loss: 0.0029


23/67 [=========>....................] - ETA: 1s - loss: 0.0026


25/67 [==========>...................] - ETA: 1s - loss: 0.0024


27/67 [===========>..................] - ETA: 1s - loss: 0.0022


30/67 [============>.................] - ETA: 0s - loss: 0.0021


33/67 [=============>................] - ETA: 0s - loss: 0.0020


36/67 [===============>..............] - ETA: 0s - loss: 0.0019


38/67 [================>.............] - ETA: 0s - loss: 0.0018


40/67 [================>.............] - ETA: 0s - loss: 0.0017


42/67 [=================>............] - ETA: 0s - loss: 0.0017


44/67 [==================>...........] - ETA: 0s - loss: 0.0016


47/67 [====================>.........] - ETA: 0s - loss: 0.0015


49/67 [====================>.........] - ETA: 0s - loss: 0.0015


52/67 [======================>.......] - ETA: 0s - loss: 0.0015


54/67 [=======================>......] - ETA: 0s - loss: 0.0015


57/67 [========================>.....] - ETA: 0s - loss: 0.0014


60/67 [=========================>....] - ETA: 0s - loss: 0.0014


62/67 [==========================>...] - ETA: 0s - loss: 0.0014


65/67 [============================>.] - ETA: 0s - loss: 0.0014


67/67 [==============================] - 2s 27ms/step - loss: 0.0014 - val_loss: 9.5578e-04

Epoch 5/40

 1/67 [..............................] - ETA: 2s - loss: 5.9994e-04


 3/67 [>.............................] - ETA: 2s - loss: 6.7950e-04


 6/67 [=>............................] - ETA: 1s - loss: 6.3105e-04


 9/67 [===>..........................] - ETA: 1s - loss: 5.7106e-04


12/67 [====>.........................] - ETA: 1s - loss: 5.0571e-04


15/67 [=====>........................] - ETA: 1s - loss: 5.4580e-04


18/67 [=======>......................] - ETA: 1s - loss: 4.7380e-04


20/67 [=======>......................] - ETA: 1s - loss: 4.6475e-04


23/67 [=========>....................] - ETA: 1s - loss: 4.1883e-04


26/67 [==========>...................] - ETA: 1s - loss: 4.0039e-04


29/67 [===========>..................] - ETA: 0s - loss: 3.7408e-04


32/67 [=============>................] - ETA: 0s - loss: 3.5186e-04


34/67 [==============>...............] - ETA: 0s - loss: 3.3919e-04


36/67 [===============>..............] - ETA: 0s - loss: 3.2951e-04


38/67 [================>.............] - ETA: 0s - loss: 3.2088e-04


40/67 [================>.............] - ETA: 0s - loss: 3.2164e-04


42/67 [=================>............] - ETA: 0s - loss: 3.2235e-04


44/67 [==================>...........] - ETA: 0s - loss: 3.1928e-04


46/67 [===================>..........] - ETA: 0s - loss: 3.1957e-04


48/67 [====================>.........] - ETA: 0s - loss: 3.4922e-04


50/67 [=====================>........] - ETA: 0s - loss: 3.4621e-04


53/67 [======================>.......] - ETA: 0s - loss: 3.8612e-04


56/67 [========================>.....] - ETA: 0s - loss: 3.9892e-04


59/67 [=========================>....] - ETA: 0s - loss: 4.2692e-04


62/67 [==========================>...] - ETA: 0s - loss: 4.3815e-04


65/67 [============================>.] - ETA: 0s - loss: 4.3396e-04


67/67 [==============================] - 2s 27ms/step - loss: 4.3329e-04 - val_loss: 9.3428e-04

Epoch 6/40

 1/67 [..............................] - ETA: 2s - loss: 1.0477e-04


 3/67 [>.............................] - ETA: 1s - loss: 9.4719e-05


 5/67 [=>............................] - ETA: 1s - loss: 8.6982e-05


 7/67 [==>...........................] - ETA: 1s - loss: 7.8438e-05


 9/67 [===>..........................] - ETA: 1s - loss: 7.1693e-05


12/67 [====>.........................] - ETA: 1s - loss: 7.1039e-05


14/67 [=====>........................] - ETA: 1s - loss: 8.5230e-05


16/67 [======>.......................] - ETA: 1s - loss: 1.1507e-04


19/67 [=======>......................] - ETA: 1s - loss: 1.1286e-04


22/67 [========>.....................] - ETA: 1s - loss: 1.3538e-04


25/67 [==========>...................] - ETA: 1s - loss: 1.2495e-04


27/67 [===========>..................] - ETA: 1s - loss: 1.1886e-04


29/67 [===========>..................] - ETA: 0s - loss: 1.1411e-04


31/67 [============>.................] - ETA: 0s - loss: 1.1201e-04


34/67 [==============>...............] - ETA: 0s - loss: 1.1124e-04


37/67 [===============>..............] - ETA: 0s - loss: 1.1080e-04


39/67 [================>.............] - ETA: 0s - loss: 1.0950e-04


42/67 [=================>............] - ETA: 0s - loss: 1.1665e-04


44/67 [==================>...........] - ETA: 0s - loss: 1.2073e-04


47/67 [====================>.........] - ETA: 0s - loss: 1.4811e-04


50/67 [=====================>........] - ETA: 0s - loss: 1.6560e-04


52/67 [======================>.......] - ETA: 0s - loss: 2.0732e-04


55/67 [=======================>......] - ETA: 0s - loss: 2.3088e-04


58/67 [========================>.....] - ETA: 0s - loss: 2.6143e-04


60/67 [=========================>....] - ETA: 0s - loss: 2.8093e-04


63/67 [===========================>..] - ETA: 0s - loss: 2.9059e-04


65/67 [============================>.] - ETA: 0s - loss: 2.9992e-04


67/67 [==============================] - 2s 27ms/step - loss: 2.9866e-04 - val_loss: 8.7440e-04

Epoch 7/40

 1/67 [..............................] - ETA: 1s - loss: 9.3107e-05


 4/67 [>.............................] - ETA: 1s - loss: 1.1192e-04


 7/67 [==>...........................] - ETA: 1s - loss: 1.1373e-04


10/67 [===>..........................] - ETA: 1s - loss: 9.4921e-05


13/67 [====>.........................] - ETA: 1s - loss: 1.1240e-04


16/67 [======>.......................] - ETA: 1s - loss: 1.6556e-04


19/67 [=======>......................] - ETA: 1s - loss: 1.5737e-04


22/67 [========>.....................] - ETA: 1s - loss: 1.6835e-04


24/67 [=========>....................] - ETA: 1s - loss: 1.7280e-04


26/67 [==========>...................] - ETA: 1s - loss: 1.7158e-04


28/67 [===========>..................] - ETA: 1s - loss: 1.6747e-04


31/67 [============>.................] - ETA: 0s - loss: 1.5802e-04


34/67 [==============>...............] - ETA: 0s - loss: 1.5602e-04


37/67 [===============>..............] - ETA: 0s - loss: 1.5477e-04


40/67 [================>.............] - ETA: 0s - loss: 1.5836e-04


42/67 [=================>............] - ETA: 0s - loss: 1.6479e-04


44/67 [==================>...........] - ETA: 0s - loss: 1.6640e-04


47/67 [====================>.........] - ETA: 0s - loss: 1.9908e-04


50/67 [=====================>........] - ETA: 0s - loss: 2.1162e-04


52/67 [======================>.......] - ETA: 0s - loss: 2.5294e-04


55/67 [=======================>......] - ETA: 0s - loss: 2.7303e-04


58/67 [========================>.....] - ETA: 0s - loss: 3.0660e-04


61/67 [==========================>...] - ETA: 0s - loss: 3.3013e-04


64/67 [===========================>..] - ETA: 0s - loss: 3.3269e-04


67/67 [==============================] - ETA: 0s - loss: 3.3066e-04


67/67 [==============================] - 2s 27ms/step - loss: 3.3066e-04 - val_loss: 8.5611e-04

Epoch 8/40

 1/67 [..............................] - ETA: 2s - loss: 2.8977e-05


 4/67 [>.............................] - ETA: 1s - loss: 2.3987e-05


 7/67 [==>...........................] - ETA: 1s - loss: 2.3057e-05


10/67 [===>..........................] - ETA: 1s - loss: 2.0832e-05


13/67 [====>.........................] - ETA: 1s - loss: 4.3473e-05


16/67 [======>.......................] - ETA: 1s - loss: 9.8491e-05


18/67 [=======>......................] - ETA: 1s - loss: 1.0009e-04


20/67 [=======>......................] - ETA: 1s - loss: 1.3012e-04


22/67 [========>.....................] - ETA: 1s - loss: 1.2218e-04


24/67 [=========>....................] - ETA: 1s - loss: 1.2018e-04


26/67 [==========>...................] - ETA: 1s - loss: 1.1704e-04


28/67 [===========>..................] - ETA: 1s - loss: 1.1406e-04


30/67 [============>.................] - ETA: 1s - loss: 1.1001e-04


33/67 [=============>................] - ETA: 0s - loss: 1.0953e-04


35/67 [==============>...............] - ETA: 0s - loss: 1.1079e-04


37/67 [===============>..............] - ETA: 0s - loss: 1.0919e-04


40/67 [================>.............] - ETA: 0s - loss: 1.1011e-04


43/67 [==================>...........] - ETA: 0s - loss: 1.1938e-04


46/67 [===================>..........] - ETA: 0s - loss: 1.2963e-04


48/67 [====================>.........] - ETA: 0s - loss: 1.6458e-04


50/67 [=====================>........] - ETA: 0s - loss: 1.6708e-04


52/67 [======================>.......] - ETA: 0s - loss: 2.0891e-04


54/67 [=======================>......] - ETA: 0s - loss: 2.3417e-04


57/67 [========================>.....] - ETA: 0s - loss: 2.4108e-04


59/67 [=========================>....] - ETA: 0s - loss: 2.7059e-04


62/67 [==========================>...] - ETA: 0s - loss: 2.9560e-04


64/67 [===========================>..] - ETA: 0s - loss: 3.0037e-04


66/67 [============================>.] - ETA: 0s - loss: 2.9805e-04


67/67 [==============================] - 2s 30ms/step - loss: 2.9848e-04 - val_loss: 8.2353e-04

Epoch 9/40

 1/67 [..............................] - ETA: 3s - loss: 3.6394e-05


 3/67 [>.............................] - ETA: 2s - loss: 5.5096e-05


 5/67 [=>............................] - ETA: 1s - loss: 4.4846e-05


 7/67 [==>...........................] - ETA: 1s - loss: 5.5314e-05


 9/67 [===>..........................] - ETA: 1s - loss: 4.9586e-05


11/67 [===>..........................] - ETA: 1s - loss: 6.0726e-05


13/67 [====>.........................] - ETA: 1s - loss: 7.4846e-05


16/67 [======>.......................] - ETA: 1s - loss: 1.3470e-04


18/67 [=======>......................] - ETA: 1s - loss: 1.3259e-04


21/67 [========>.....................] - ETA: 1s - loss: 1.4900e-04


23/67 [=========>....................] - ETA: 1s - loss: 1.4589e-04


25/67 [==========>...................] - ETA: 1s - loss: 1.5582e-04


27/67 [===========>..................] - ETA: 1s - loss: 1.5346e-04


29/67 [===========>..................] - ETA: 1s - loss: 1.4771e-04


32/67 [=============>................] - ETA: 1s - loss: 1.4084e-04


34/67 [==============>...............] - ETA: 0s - loss: 1.4329e-04


37/67 [===============>..............] - ETA: 0s - loss: 1.4294e-04


39/67 [================>.............] - ETA: 0s - loss: 1.4445e-04


41/67 [=================>............] - ETA: 0s - loss: 1.4637e-04


44/67 [==================>...........] - ETA: 0s - loss: 1.5376e-04


46/67 [===================>..........] - ETA: 0s - loss: 1.6308e-04


49/67 [====================>.........] - ETA: 0s - loss: 2.0130e-04


51/67 [=====================>........] - ETA: 0s - loss: 2.0347e-04


53/67 [======================>.......] - ETA: 0s - loss: 2.5942e-04


55/67 [=======================>......] - ETA: 0s - loss: 2.6981e-04


58/67 [========================>.....] - ETA: 0s - loss: 3.0039e-04


60/67 [=========================>....] - ETA: 0s - loss: 3.1883e-04


62/67 [==========================>...] - ETA: 0s - loss: 3.2908e-04


64/67 [===========================>..] - ETA: 0s - loss: 3.3176e-04


66/67 [============================>.] - ETA: 0s - loss: 3.2879e-04


67/67 [==============================] - 2s 32ms/step - loss: 3.2923e-04 - val_loss: 8.0485e-04

Epoch 10/40

 1/67 [..............................] - ETA: 2s - loss: 1.8117e-05


 3/67 [>.............................] - ETA: 2s - loss: 2.4425e-05


 5/67 [=>............................] - ETA: 1s - loss: 1.8337e-05


 7/67 [==>...........................] - ETA: 2s - loss: 2.6269e-05


 9/67 [===>..........................] - ETA: 2s - loss: 2.4380e-05


11/67 [===>..........................] - ETA: 2s - loss: 3.8029e-05


13/67 [====>.........................] - ETA: 2s - loss: 5.5378e-05


15/67 [=====>........................] - ETA: 2s - loss: 1.2398e-04


17/67 [======>.......................] - ETA: 1s - loss: 1.1694e-04


18/67 [=======>......................] - ETA: 2s - loss: 1.1927e-04


19/67 [=======>......................] - ETA: 2s - loss: 1.1925e-04


21/67 [========>.....................] - ETA: 1s - loss: 1.3914e-04


23/67 [=========>....................] - ETA: 1s - loss: 1.3461e-04


25/67 [==========>...................] - ETA: 1s - loss: 1.4239e-04


27/67 [===========>..................] - ETA: 1s - loss: 1.4013e-04


29/67 [===========>..................] - ETA: 1s - loss: 1.3527e-04


32/67 [=============>................] - ETA: 1s - loss: 1.2991e-04


35/67 [==============>...............] - ETA: 1s - loss: 1.3277e-04


37/67 [===============>..............] - ETA: 1s - loss: 1.3056e-04


39/67 [================>.............] - ETA: 0s - loss: 1.3123e-04


41/67 [=================>............] - ETA: 0s - loss: 1.3315e-04


44/67 [==================>...........] - ETA: 0s - loss: 1.4125e-04


47/67 [====================>.........] - ETA: 0s - loss: 1.7760e-04


49/67 [====================>.........] - ETA: 0s - loss: 1.8796e-04


51/67 [=====================>........] - ETA: 0s - loss: 1.9064e-04


53/67 [======================>.......] - ETA: 0s - loss: 2.4738e-04


55/67 [=======================>......] - ETA: 0s - loss: 2.5958e-04


57/67 [========================>.....] - ETA: 0s - loss: 2.6489e-04


59/67 [=========================>....] - ETA: 0s - loss: 2.9414e-04


61/67 [==========================>...] - ETA: 0s - loss: 3.1871e-04


63/67 [===========================>..] - ETA: 0s - loss: 3.1572e-04


65/67 [============================>.] - ETA: 0s - loss: 3.2186e-04


67/67 [==============================] - 2s 34ms/step - loss: 3.2093e-04 - val_loss: 7.8456e-04

Epoch 11/40

 1/67 [..............................] - ETA: 1s - loss: 2.4023e-05


 3/67 [>.............................] - ETA: 1s - loss: 3.7469e-05


 5/67 [=>............................] - ETA: 1s - loss: 3.0589e-05


 8/67 [==>...........................] - ETA: 1s - loss: 4.1178e-05


11/67 [===>..........................] - ETA: 1s - loss: 5.4490e-05


13/67 [====>.........................] - ETA: 1s - loss: 7.2664e-05


15/67 [=====>........................] - ETA: 1s - loss: 1.4345e-04


18/67 [=======>......................] - ETA: 1s - loss: 1.3572e-04


21/67 [========>.....................] - ETA: 1s - loss: 1.5108e-04


23/67 [=========>....................] - ETA: 1s - loss: 1.5182e-04


26/67 [==========>...................] - ETA: 1s - loss: 1.6534e-04


29/67 [===========>..................] - ETA: 0s - loss: 1.5703e-04


32/67 [=============>................] - ETA: 0s - loss: 1.4887e-04


34/67 [==============>...............] - ETA: 0s - loss: 1.5255e-04


37/67 [===============>..............] - ETA: 0s - loss: 1.5200e-04


40/67 [================>.............] - ETA: 0s - loss: 1.5226e-04


42/67 [=================>............] - ETA: 0s - loss: 1.5852e-04


45/67 [===================>..........] - ETA: 0s - loss: 1.6066e-04


47/67 [====================>.........] - ETA: 0s - loss: 1.9989e-04


49/67 [====================>.........] - ETA: 0s - loss: 2.0874e-04


52/67 [======================>.......] - ETA: 0s - loss: 2.5282e-04


54/67 [=======================>......] - ETA: 0s - loss: 2.8255e-04


57/67 [========================>.....] - ETA: 0s - loss: 2.8671e-04


59/67 [=========================>....] - ETA: 0s - loss: 3.1653e-04


62/67 [==========================>...] - ETA: 0s - loss: 3.4280e-04


65/67 [============================>.] - ETA: 0s - loss: 3.4363e-04


67/67 [==============================] - 2s 27ms/step - loss: 3.4252e-04 - val_loss: 7.7259e-04

Epoch 12/40

 1/67 [..............................] - ETA: 2s - loss: 1.9856e-05


 4/67 [>.............................] - ETA: 1s - loss: 2.6365e-05


 6/67 [=>............................] - ETA: 1s - loss: 3.2153e-05


 8/67 [==>...........................] - ETA: 1s - loss: 3.6681e-05


10/67 [===>..........................] - ETA: 1s - loss: 3.6016e-05


12/67 [====>.........................] - ETA: 1s - loss: 5.0130e-05


14/67 [=====>........................] - ETA: 1s - loss: 9.3710e-05


16/67 [======>.......................] - ETA: 1s - loss: 1.4004e-04


18/67 [=======>......................] - ETA: 1s - loss: 1.3790e-04


20/67 [=======>......................] - ETA: 1s - loss: 1.5498e-04


23/67 [=========>....................] - ETA: 1s - loss: 1.5504e-04


26/67 [==========>...................] - ETA: 1s - loss: 1.7017e-04


29/67 [===========>..................] - ETA: 1s - loss: 1.6141e-04


32/67 [=============>................] - ETA: 0s - loss: 1.5274e-04


34/67 [==============>...............] - ETA: 0s - loss: 1.5665e-04


36/67 [===============>..............] - ETA: 0s - loss: 1.5631e-04


38/67 [================>.............] - ETA: 0s - loss: 1.5396e-04


41/67 [=================>............] - ETA: 0s - loss: 1.5540e-04


43/67 [==================>...........] - ETA: 0s - loss: 1.6179e-04


46/67 [===================>..........] - ETA: 0s - loss: 1.7309e-04


49/67 [====================>.........] - ETA: 0s - loss: 2.1154e-04


51/67 [=====================>........] - ETA: 0s - loss: 2.1256e-04


54/67 [=======================>......] - ETA: 0s - loss: 2.8735e-04


56/67 [========================>.....] - ETA: 0s - loss: 2.9324e-04


58/67 [========================>.....] - ETA: 0s - loss: 3.1390e-04


61/67 [==========================>...] - ETA: 0s - loss: 3.4682e-04


64/67 [===========================>..] - ETA: 0s - loss: 3.5101e-04


67/67 [==============================] - ETA: 0s - loss: 3.4771e-04


67/67 [==============================] - 2s 28ms/step - loss: 3.4771e-04 - val_loss: 7.5997e-04

Epoch 13/40

 1/67 [..............................] - ETA: 1s - loss: 2.2567e-05


 3/67 [>.............................] - ETA: 2s - loss: 3.5764e-05


 5/67 [=>............................] - ETA: 1s - loss: 3.0406e-05


 7/67 [==>...........................] - ETA: 1s - loss: 4.5067e-05


 8/67 [==>...........................] - ETA: 2s - loss: 4.4225e-05


10/67 [===>..........................] - ETA: 1s - loss: 4.3376e-05


12/67 [====>.........................] - ETA: 1s - loss: 5.8485e-05


14/67 [=====>........................] - ETA: 1s - loss: 1.0550e-04


16/67 [======>.......................] - ETA: 1s - loss: 1.5171e-04


18/67 [=======>......................] - ETA: 1s - loss: 1.4820e-04


20/67 [=======>......................] - ETA: 1s - loss: 1.6201e-04


22/67 [========>.....................] - ETA: 1s - loss: 1.6291e-04


25/67 [==========>...................] - ETA: 1s - loss: 1.8812e-04


27/67 [===========>..................] - ETA: 1s - loss: 1.8526e-04


29/67 [===========>..................] - ETA: 1s - loss: 1.7705e-04


32/67 [=============>................] - ETA: 1s - loss: 1.6693e-04


34/67 [==============>...............] - ETA: 1s - loss: 1.7273e-04


36/67 [===============>..............] - ETA: 0s - loss: 1.7234e-04


38/67 [================>.............] - ETA: 0s - loss: 1.6975e-04


41/67 [=================>............] - ETA: 0s - loss: 1.6926e-04


43/67 [==================>...........] - ETA: 0s - loss: 1.7582e-04


46/67 [===================>..........] - ETA: 0s - loss: 1.8756e-04


48/67 [====================>.........] - ETA: 0s - loss: 2.2628e-04


51/67 [=====================>........] - ETA: 0s - loss: 2.2621e-04


52/67 [======================>.......] - ETA: 0s - loss: 2.7051e-04


54/67 [=======================>......] - ETA: 0s - loss: 3.0324e-04


57/67 [========================>.....] - ETA: 0s - loss: 3.0597e-04


59/67 [=========================>....] - ETA: 0s - loss: 3.3523e-04


62/67 [==========================>...] - ETA: 0s - loss: 3.6389e-04


65/67 [============================>.] - ETA: 0s - loss: 3.6447e-04


67/67 [==============================] - ETA: 0s - loss: 3.6286e-04


67/67 [==============================] - 2s 31ms/step - loss: 3.6286e-04 - val_loss: 7.5029e-04

Epoch 14/40

 1/67 [..............................] - ETA: 1s - loss: 2.0981e-05


 4/67 [>.............................] - ETA: 1s - loss: 2.9686e-05

plt.plot(history.history['loss'], label='train')
plt.plot(history.history['val_loss'], label='test')
plt.legend();
../../_images/rnr_accion_Apple_Prediccion_tres_dias-multiple_34_0.png

Predicciones#

Prepara los datos de validación#

X_test.shape

(534, 60)

Calcula predicciones#

# predictions
prediction = model_01.predict(X_test)
#prediction = scaler.inverse_transform(prediction)

17/17 [==============================] - 1s 15ms/step

print(X_test.shape)
print(prediction.shape)
print(y_test.shape)

(534, 60)
(534, 10)
(534, 10)

Prepara datos para mostrar predicciones#

pred =0
y_train_p = X_test[pred,:]
y_test_p = y_test[pred,:]
y_pred_p = prediction[pred,:]

#print(y_train_p.shape)
#print(y_test_p.shape)
#print(y_pred_p.shape)

Gráfica de las predicciones#

plt.plot(np.arange(0, len(y_train_p)), y_train_p, 'g', label="historia")
plt.plot(np.arange(len(y_train_p), len(y_train_p) + len(y_test_p)), y_test_p, marker='.', label="verdadero")
plt.plot(np.arange(len(y_train_p), len(y_train_p) + len(y_test_p)), y_pred_p, 'r', label="predicción")
#plt.ylabel('Valor')
plt.xlabel('Time Step')
plt.title("Apple: Historia del precio la acción al cierre. Escala (0,1)", size = 20)
plt.legend()
plt.show();
../../_images/rnr_accion_Apple_Prediccion_tres_dias-multiple_44_0.png

Guarda el modelo entrenado#

model_01.save('../Datos/modelo_Apple_3_dia_multipl3.h5')

Intervalos de confianza. TO DO#

model_01.get_config()

{'name': 'series_LSTM_model',
 'layers': [{'class_name': 'InputLayer',
   'config': {'batch_input_shape': (None, 60, 1),
    'dtype': 'float32',
    'sparse': False,
    'ragged': False,
    'name': 'input_1'},
   'name': 'input_1',
   'inbound_nodes': []},
  {'class_name': 'Dropout',
   'config': {'name': 'Dropout_01',
    'trainable': True,
    'dtype': 'float32',
    'rate': 0.0,
    'noise_shape': None,
    'seed': None},
   'name': 'Dropout_01',
   'inbound_nodes': [[['input_1', 0, 0, {}]]]},
  {'class_name': 'LSTM',
   'config': {'name': 'LSTM_layer',
    'trainable': True,
    'dtype': 'float32',
    'return_sequences': False,
    'return_state': False,
    'go_backwards': False,
    'stateful': False,
    'unroll': False,
    'time_major': False,
    'units': 50,
    'activation': 'tanh',
    'recurrent_activation': 'sigmoid',
    'use_bias': True,
    'kernel_initializer': {'class_name': 'GlorotUniform',
     'config': {'seed': None},
     'shared_object_id': 2},
    'recurrent_initializer': {'class_name': 'Orthogonal',
     'config': {'gain': 1.0, 'seed': None},
     'shared_object_id': 3},
    'bias_initializer': {'class_name': 'Zeros',
     'config': {},
     'shared_object_id': 4},
    'unit_forget_bias': True,
    'kernel_regularizer': None,
    'recurrent_regularizer': None,
    'bias_regularizer': None,
    'activity_regularizer': None,
    'kernel_constraint': None,
    'recurrent_constraint': None,
    'bias_constraint': None,
    'dropout': 0.0,
    'recurrent_dropout': 0.0,
    'implementation': 2},
   'name': 'LSTM_layer',
   'inbound_nodes': [[['Dropout_01', 0, 0, {}]]]},
  {'class_name': 'Dropout',
   'config': {'name': 'Dropout_02',
    'trainable': True,
    'dtype': 'float32',
    'rate': 0.0,
    'noise_shape': None,
    'seed': None},
   'name': 'Dropout_02',
   'inbound_nodes': [[['LSTM_layer', 0, 0, {}]]]},
  {'class_name': 'Dense',
   'config': {'name': 'dense',
    'trainable': True,
    'dtype': 'float32',
    'units': 10,
    'activation': 'linear',
    'use_bias': True,
    'kernel_initializer': {'class_name': 'GlorotUniform',
     'config': {'seed': None}},
    'bias_initializer': {'class_name': 'Zeros', 'config': {}},
    'kernel_regularizer': None,
    'bias_regularizer': None,
    'activity_regularizer': None,
    'kernel_constraint': None,
    'bias_constraint': None},
   'name': 'dense',
   'inbound_nodes': [[['Dropout_02', 0, 0, {}]]]}],
 'input_layers': [['input_1', 0, 0]],
 'output_layers': [['dense', 0, 0]]}

Referencias#

  1. Introducción a Redes LSTM

  2. Time Series Forecasting with LSTMs using TensorFlow 2 and Keras in Python

  3. Dive into Deep Learnig

  4. Understanding LSTM Networks

  5. Ralf C. Staudemeyer and Eric Rothstein Morris,Understanding LSTM a tutorial into Long Short-Term Memory Recurrent Neural Networks, arxiv, September 2019

  6. Karpathy, The Unreasonable Effectiveness of Recurrent Neural Networks

  7. Anton Lucanus, Making Automation More Efficient by Learning from Historical Trade Data, 8:43 AM, January 7, 2020

  8. https://www.analyticsvidhya.com/blog/2018/10/predicting-stock-price-machine-learningnd-deep-learning-techniques-python/

  9. https://www.youtube.com/watch?v=2BrpKpWwT2A&list=PLQVvvaa0QuDcOdF96TBtRtuQksErCEBYZ&index=1

  10. https://towardsdatascience.com/using-lstms-for-stock-market-predictions-tensorflow-9e83999d4653

  11. https://github.com/llSourcell/Reinforcement_Learning_for_Stock_Prediction/blob/master/README.md