Google Stock Price Prediction

import math
import pandas as pd
import datetime
import pandas_datareader as web
import numpy as np
import seaborn as sns
import statsmodels.api as sm
from sklearn.preprocessing import MinMaxScaler
from sklearn.tree import DecisionTreeRegressor
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, LSTM, Bidirectional, Dropout, Activation, GRU
from plotly.graph_objs import *
from plotly.offline import init_notebook_mode, iplot, iplot_mpl
import matplotlib
import matplotlib.pyplot as plt
matplotlib.style.use('seaborn')

# Importing the training set
data = pd.read_csv('Data/googl.csv', index_col="Date",parse_dates=True)
data

data.shape

data1 = data.filter(['Close'])

training_set = data1.values
training_data_len = math.ceil( len(training_set) * .8 )

# Feature Scaling
sc = MinMaxScaler(feature_range = (0, 1))
training_set_scaled = sc.fit_transform(training_set)

# Creating a data structure with 60 timesteps and 1 output
X_train_lstm = []
y_train_lstm = []
for i in range(60, 1361):
    X_train_lstm.append(training_set_scaled[i-60:i, 0])
    y_train_lstm.append(training_set_scaled[i, 0])
X_train_lstm, y_train_lstm = np.array(X_train_lstm), np.array(y_train_lstm)

# Reshaping
X_train_lstm = np.reshape(X_train_lstm, (X_train_lstm.shape[0], X_train_lstm.shape[1], 1))

# Initialising the RNN
regressor = Sequential()

# Adding the first LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50, return_sequences = True, input_shape = (X_train_lstm.shape[1], 1)))
regressor.add(Dropout(0.2))

# Adding a second LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50, return_sequences = True))
regressor.add(Dropout(0.2))

# Adding a third LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50, return_sequences = True))
regressor.add(Dropout(0.2))

# Adding a fourth LSTM layer and some Dropout regularisation
regressor.add(LSTM(units = 50))
regressor.add(Dropout(0.2))

# Adding the output layer
regressor.add(Dense(units = 1))

# Compiling the RNN
regressor.compile(optimizer = 'adam', loss = 'mean_squared_error')

# Fitting the RNN to the Training set
regressor.fit(X_train_lstm, y_train_lstm, epochs = 100, batch_size = 32)

Epoch 1/100
41/41 [==============================] - 9s 231ms/step - loss: 0.0276
Epoch 2/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0057
Epoch 3/100
41/41 [==============================] - 9s 208ms/step - loss: 0.0060
Epoch 4/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0054
Epoch 5/100
41/41 [==============================] - 9s 211ms/step - loss: 0.0056
Epoch 6/100
41/41 [==============================] - 9s 208ms/step - loss: 0.0043
Epoch 7/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0045
Epoch 8/100
41/41 [==============================] - 9s 211ms/step - loss: 0.0039
Epoch 9/100
41/41 [==============================] - 9s 208ms/step - loss: 0.0039
Epoch 10/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0041
Epoch 11/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0041
Epoch 12/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0040
Epoch 13/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0037
Epoch 14/100
41/41 [==============================] - 9s 209ms/step - loss: 0.0032
Epoch 15/100
41/41 [==============================] - 9s 211ms/step - loss: 0.0039
Epoch 16/100
41/41 [==============================] - 9s 211ms/step - loss: 0.0032
Epoch 17/100
41/41 [==============================] - 9s 211ms/step - loss: 0.0030
Epoch 18/100
41/41 [==============================] - 9s 213ms/step - loss: 0.0030
Epoch 19/100
41/41 [==============================] - 9s 214ms/step - loss: 0.0027
Epoch 20/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0033
Epoch 21/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0032
Epoch 22/100
41/41 [==============================] - 9s 227ms/step - loss: 0.0031
Epoch 23/100
41/41 [==============================] - 9s 214ms/step - loss: 0.0028
Epoch 24/100
41/41 [==============================] - 9s 213ms/step - loss: 0.0029
Epoch 25/100
41/41 [==============================] - 9s 226ms/step - loss: 0.0029
Epoch 26/100
41/41 [==============================] - 10s 246ms/step - loss: 0.0025
Epoch 27/100
41/41 [==============================] - 9s 228ms/step - loss: 0.0024
Epoch 28/100
41/41 [==============================] - 10s 243ms/step - loss: 0.0024
Epoch 29/100
41/41 [==============================] - 9s 228ms/step - loss: 0.0025
Epoch 30/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0026
Epoch 31/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0026
Epoch 32/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0026
Epoch 33/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0028
Epoch 34/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0022
Epoch 35/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0026
Epoch 36/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0022
Epoch 37/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0023
Epoch 38/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0022
Epoch 39/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0023
Epoch 40/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0023
Epoch 41/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0021
Epoch 42/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0020
Epoch 43/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0020
Epoch 44/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0023
Epoch 45/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0023
Epoch 46/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0022
Epoch 47/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0023
Epoch 48/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0020
Epoch 49/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0018
Epoch 50/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0020
Epoch 51/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0021
Epoch 52/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0018
Epoch 53/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0019
Epoch 54/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0018
Epoch 55/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0017
Epoch 56/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0018
Epoch 57/100
41/41 [==============================] - 9s 220ms/step - loss: 0.0016
Epoch 58/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0018
Epoch 59/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0015
Epoch 60/100
41/41 [==============================] - 9s 230ms/step - loss: 0.0017
Epoch 61/100
41/41 [==============================] - 10s 236ms/step - loss: 0.0018
Epoch 62/100
41/41 [==============================] - 9s 230ms/step - loss: 0.0017
Epoch 63/100
41/41 [==============================] - 9s 217ms/step - loss: 0.0016
Epoch 64/100
41/41 [==============================] - 9s 217ms/step - loss: 0.0017
Epoch 65/100
41/41 [==============================] - 9s 215ms/step - loss: 0.0018
Epoch 66/100
41/41 [==============================] - 9s 218ms/step - loss: 0.0016
Epoch 67/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0017
Epoch 68/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0014
Epoch 69/100
41/41 [==============================] - 9s 218ms/step - loss: 0.0017
Epoch 70/100
41/41 [==============================] - 9s 217ms/step - loss: 0.0014
Epoch 71/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0016
Epoch 72/100
41/41 [==============================] - 9s 214ms/step - loss: 0.0017
Epoch 73/100
41/41 [==============================] - 9s 208ms/step - loss: 0.0014
Epoch 74/100
41/41 [==============================] - 9s 216ms/step - loss: 0.0014
Epoch 75/100
41/41 [==============================] - 9s 223ms/step - loss: 0.0013
Epoch 76/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0014
Epoch 77/100
41/41 [==============================] - 9s 223ms/step - loss: 0.0014
Epoch 78/100
41/41 [==============================] - 9s 224ms/step - loss: 0.0013
Epoch 79/100
41/41 [==============================] - 9s 225ms/step - loss: 0.0014
Epoch 80/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0013
Epoch 81/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0012
Epoch 82/100
41/41 [==============================] - 9s 218ms/step - loss: 0.0015
Epoch 83/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0013
Epoch 84/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0015
Epoch 85/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0012
Epoch 86/100
41/41 [==============================] - 9s 221ms/step - loss: 0.0012
Epoch 87/100
41/41 [==============================] - 9s 225ms/step - loss: 0.0013
Epoch 88/100
41/41 [==============================] - 9s 224ms/step - loss: 0.0011
Epoch 89/100
41/41 [==============================] - 9s 219ms/step - loss: 0.0012
Epoch 90/100
41/41 [==============================] - 9s 218ms/step - loss: 0.0012
Epoch 91/100
41/41 [==============================] - 9s 226ms/step - loss: 0.0013
Epoch 92/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0012
Epoch 93/100
41/41 [==============================] - 9s 223ms/step - loss: 0.0011
Epoch 94/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0013
Epoch 95/100
41/41 [==============================] - 9s 230ms/step - loss: 0.0012
Epoch 96/100
41/41 [==============================] - 9s 225ms/step - loss: 0.0012
Epoch 97/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0012
Epoch 98/100
41/41 [==============================] - 9s 226ms/step - loss: 0.0013
Epoch 99/100
41/41 [==============================] - 10s 235ms/step - loss: 0.0013
Epoch 100/100
41/41 [==============================] - 9s 222ms/step - loss: 0.0011
Out[9]:
<tensorflow.python.keras.callbacks.History at 0x14e18c780>

# Getting the predicted stock price of 2017
dataset_total = pd.concat((data['Close'], dataset_test['Close']), axis = 0)
inputs = dataset_total[len(dataset_total) - len(dataset_test) - 60:].values
inputs = inputs.reshape(-1,1)
inputs = sc.transform(inputs)
X_test_lstm = []
y_test_lstm = []
for i in range(60, 83):
    X_test_lstm.append(inputs[i-60:i, 0])
    y_test_lstm.append(inputs[i, 0])
X_test_lstm = np.array(X_test_lstm)
X_test_lstm = np.reshape(X_test_lstm, (X_test_lstm.shape[0], X_test_lstm.shape[1], 1))
predicted_stock_price = regressor.predict(X_test_lstm)
predicted_stock_price = sc.inverse_transform(predicted_stock_price)

# Visualising the results
plt.figure(figsize=(14,7))
plt.plot(real_stock_price, color = 'red', label = 'Real Google Stock Price')
plt.plot(predicted_stock_price, color = 'blue', label = 'Predicted Google Stock Price')
plt.title('Google Stock Price Prediction')
plt.xlabel('Time')
plt.ylabel('Google Stock Price')
plt.legend()
plt.show()

y_test_lstm = np.array(y_test_lstm)

#sc.scale
mean_y_test_lstm = y_test_lstm.mean()
mean_y_hat_lstm = predicted_stock_price.mean()

print(mean_y_hat_lstm, mean_y_test_lstm)

normal_scale = 1/0.00097294
normal_scale

y_test_lstm = y_test_lstm * normal_scale

accuracy = round((mean_y_test_lstm / mean_y_hat_lstm) * 100,2)
accuracy