Deep learning LSTM model for stock analysis and prediction

Project description ：

This program uses the long-term and short-term memory model in the cyclic neural network （LSTM） With Apple The company used to 60 To predict the closing price
LSTM The model is equivalent to RNN An improved version of the model
RNN： Want to remember all the information , Whether it's useful information or useless information （ Heavy burden of memory ）
LSTM： Design a memory cell , It has the function of selective memory , You can choose to remember important information , Filter out noise information , Lighten the burden of memory
About LSTM The principle of this does not need to be repeated , Recommended viewing B Stop video .
Use Anaconda3 Medium Jupyter To write the ( Personal recommendation is very easy to use )

---

Environment building ：

Anaconda install
Anaconda Add a new library

Call directly after entering pip that will do

pip install tensorflow -i https://pypi.tuna.tsinghua.edu.cn/simple --user

hold tensorflow Change to the package name you want to import

---

Project steps ：

1、 Import library first ：

# Import library 
import math
import pandas_datareader as web
import numpy as np
import pandas as pd
import seaborn as sns
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras. layers import Dense,LSTM
import matplotlib.pyplot as plt

2、 Get stock quotes ：

# Solve the font display exception bug
sns.set_style("whitegrid", {
    "font.sans-serif": ['KaiTi', 'Arial']}) 
# Get stock quotes 
df=web.DataReader("AAPL",data_source='yahoo',start='2013-01-01',end='2022-02-15')
# Display the data 
df

The display data is as follows ：

3、 Plot the historical closing price of the stock ：

# Visualize closing price history 
plt.figure(figsize=(16,8))
plt.title(' Historical price ')
plt.plot(df['Close'])
plt.xlabel(' date ',fontsize=18)
plt.ylabel(' Closing price $  ($)',fontsize=18)
plt.show()

4、 Create one that only “Close” Data set for column

# Create one that only “Close” Data set for column 
data = df.filter(['Close'])
## Convert dataset to numpy Array 
dataset = data.values

5、 Get the number of rows of the training model

training_data_len = math.ceil( len (dataset)*.8)
training_data_len

6、 Data normalization

# Data normalization 
scaler = MinMaxScaler(feature_range=(0, 1))
scaled_data = scaler.fit_transform(dataset)
scaled_data

7、 Create training data sets

# Data normalization 
# Create training data sets 
# Create a scaled training dataset train_data
train_data = scaled_data[0:training_data_len,:]
# Split data into x_train and y_train Data sets 
x_train = []
y_train = []

for i in range(60,len(train_data)):
    x_train.append(train_data[i-60:i,0])
    y_train.append(train_data[i,0])
    if i <= 61:
        print(x_train)
        print(y_train)

8、 take x_train and y_train Convert to numpy Array

x_train,y_train = np.array(x_train),np.array(y_train)

9、 Reshaped data

x_train = np.reshape(x_train,(x_train.shape[0],x_train.shape[1],1))

10、 establish LSTM Model

model = Sequential()
model.add(LSTM(50,return_sequences=True, input_shape=(x_train.shape[1],1)))
model.add(LSTM(50,return_sequences=False))
model.add(Dense(25))
model.add(Dense(1))

11、 Compile model

model.compile(optimizer='adam',loss='mean_squared_error')

12、 Training models

model.fit(x_train,y_train,batch_size=1,epochs=1)

13、 Create test data sets

# Create a new array , Include from index 1978 To 2438 Zoom value of 
test_data = scaled_data[training_data_len - 60:,:]
# Create a dataset x test and y_test
x_test = []
y_test = dataset[training_data_len:,:]
for i in range(60,len(test_data)):
    x_test.append(test_data[i-60:i,0])
# Convert data to numpy Array 
x_test = np.array(x_test)
# Reshaped data 
x_test = np.reshape(x_test,(x_test.shape[0],x_test.shape[1],1))

14、 Get the predicted price value of the model

# Get the predicted price value of the model 
predictions = model.predict(x_test)
predictions = scaler.inverse_transform(predictions) 

---

15、 Calculate the root mean square error (RMSE)

rmse = np.sqrt(np.mean( predictions - y_test)**2 )

16、 Charting

train = data[:training_data_len]
valid = data[training_data_len:]
valid['Predictions'] = predictions
# Visualization data 
plt.figure(figsize=(16,8))
plt.title('LSTM Model to predict ')
plt.xlabel(' date ', fontsize=18)
plt.ylabel(' Closing price $  ($)', fontsize=18)
plt.plot(train['Close'])
plt.plot(valid[['Close','Predictions']])
plt.legend([" Model training value ",' True value ',' Predictive value '],loc='lower right')
plt.show()