python实现LSTM股价预测

01

今天，梅朵和大家分享用Python实现LSTM股价预测，仅用于技术学习交流，不作为投资建议。

Python实现LSTM股价预测

  

实现技巧

1.导入依赖库

主要是安装相关的依赖库。本文实现的环境为：python 3.7。

# 导入依赖库import sklearnimport numpy as npimport pandas as pdimport matplotlib.pyplot as pltfrom keras.optimizers import Adamfrom keras.models import Sequentialfrom keras.layers import LSTM, Densefrom sklearn.preprocessing import MinMaxScaler

2.读取股价数据并归一化

读取股价数据文件，并对数据进行归一化处理。

    # 读取股票价格数据    data = pd.read_csv('stock_data.csv')    prices = data['Close'].values
    # 数据预处理：归一化    scaler = MinMaxScaler()    prices = scaler.fit_transform(prices.reshape(-1, 1))

3.划分训练集和测试集并确定模型输入和输出数据

划分数据的训练集和测试集，并确定模型输入库数据和输出数据。

# 划分数据集为训练集和测试集 train_size = int(len(prices) * 0.75) train_data, test_data = prices[:train_size], prices[train_size:]
#确定模型输入和输出数据 in_out_length = 12 train_sequences = [] for i in range(len(train_data) - in_out_length): train_sequences.append(train_data[i:i+in_out_length]) X_train = np.array(train_sequences) y_train = train_data[in_out_length:] test_sequences = [] for i in range(len(test_data) - in_out_length): test_sequences.append(test_data[i:i+in_out_length]) X_test = np.array(test_sequences) y_test = test_data[in_out_length:]

4.创建LSTM模型并训练模型

创建LSTM模型并进行模型训练。

    # 创建LSTM模型    model = Sequential()    model.add(LSTM(50, input_shape=(in_out_length, 1)))    model.add(Dense(1))    learning_rate=0.002    model.compile(loss='mean_squared_error', optimizer=Adam(learning_rate))
    # 训练模型    model.fit(X_train, y_train, epochs=50, batch_size=64)

3.利用LSTM模型预测股价

利用训练好的LSTM模型进行预测。

# 预测未来股价 predicted_prices = model.predict(X_test) predicted_prices = scaler.inverse_transform(predicted_prices) y_test = scaler.inverse_transform(y_test)

完整源代码

# 导入依赖库import sklearnimport numpy as npimport pandas as pdimport matplotlib.pyplot as pltfrom keras.optimizers import Adamfrom keras.models import Sequentialfrom keras.layers import LSTM, Densefrom sklearn.preprocessing import MinMaxScaler
if __name__=='__main__':    # 读取股票价格数据    data = pd.read_csv('stock_data.csv')    prices = data['Close'].values
    # 数据预处理：归一化    scaler = MinMaxScaler()    prices = scaler.fit_transform(prices.reshape(-1, 1))
    # 划分数据集为训练集和测试集    train_size = int(len(prices) * 0.75)    train_data, test_data = prices[:train_size], prices[train_size:]
    #确定模型输入和输出数据    in_out_length = 12    train_sequences = []    for i in range(len(train_data) - in_out_length):        train_sequences.append(train_data[i:i+in_out_length])    X_train = np.array(train_sequences)    y_train = train_data[in_out_length:]    test_sequences = []    for i in range(len(test_data) - in_out_length):        test_sequences.append(test_data[i:i+in_out_length])    X_test = np.array(test_sequences)    y_test = test_data[in_out_length:]
    # 创建LSTM模型    model = Sequential()    model.add(LSTM(50, input_shape=(in_out_length, 1)))    model.add(Dense(1))    learning_rate=0.002    model.compile(loss='mean_squared_error', optimizer=Adam(learning_rate))
    # 训练模型    model.fit(X_train, y_train, epochs=50, batch_size=64)
    # 预测未来股价    predicted_prices = model.predict(X_test)    predicted_prices = scaler.inverse_transform(predicted_prices)    y_test = scaler.inverse_transform(y_test)
    # 绘制预测结果    plt.plot(y_test, label='True_Prices')    plt.plot(predicted_prices, label='Predicted_Prices')    plt.legend()    plt.xlabel('Time_Steps')    plt.ylabel('Stock_Price')    plt.title('LSTM Stock Price Prediction')    plt.show()