基于tushare数据多维度时间序列lstm预测股票价格

首先，介绍tushare，这是一个非常友好的数据源
网址是：https://tushare.pro

数据丰富
拥有丰富的数据内容，如股票、基金、期货、数字货币等行情数据，公司财务、基金经理等基本面数据

获取简单
SDK开发包支持语言，同时提供HTTP Restful接口，最大程度方便不同人群的使用

落地方便
提供多种数据储存方式，如Oracle、MySQL，MongoDB、HDF5、CSV等，为数据获取提供了性能保证

包含以下几大类的数据

悠远的发展历史

强大的合作伙伴

丰富的种类

项目设计

data:数据爬取下来的存放路径
login:各种key存储的地方
model:模型存放的地方
picture:画图存放的地方
spider:爬取数据的地方

ts.py
这一块没什么好说的，基本上的tushare上有介绍使用的方法
网址:https://tushare.pro/document/2?doc_id=27

# from login.mysql_login import engine
# CONN = engine.connect()
from login.login_sdk import pro
import pandas as pd
if __name__ == '__main__':
    ts_code = '000001.SZ'
    start_date = '20120101'
    end_date = '20191231'
    data = pro.daily(ts_code=ts_code, start_date=start_date, end_date=end_date)
    data.to_csv('../data/data.csv',index=False)

login：这里涉及到证书问题，所以就不贴代码了，详情可以到https://tushare.pro查看

model:

import pandas as pd
import numpy as np
from sklearn.preprocessing import MinMaxScaler

def create_dataset(X,Y,window_size = 30,predict_size = 5):
    data_X, data_Y = [], []
    for i in range(len(X) - window_size - predict_size + 1):
        a = X[i:(i + window_size)]
        data_X.append(a)
        data_Y.append(Y[i + window_size :i + window_size + predict_size,0])
    return(np.array(data_X), np.array(data_Y))


def minMaxScaler(df,feature_range=(0, 1)):
    scaler = MinMaxScaler(feature_range=feature_range)
    data_raw = pd.DataFrame(df).values.astype("float32")
    scaler.fit(data_raw)
    return scaler

if __name__ == '__main__':
    data = pd.read_csv('./data/data.csv')
    data = data.sort_values('trade_date').reset_index(drop=True)

    temp_data = data.copy()
    # #####################################数据预处理###############################
    import joblib
    scaler = minMaxScaler(temp_data.vol)
    temp_data.vol = \
    scaler.transform(pd.DataFrame(temp_data.vol).values.astype("float32"))
	joblib.dump(scaler, 'vol.save')  # save
    scaler = joblib.load('vol.save') # load

    scaler = minMaxScaler(temp_data.turnover_rate)
    temp_data.turnover_rate = \
    scaler.transform(pd.DataFrame(temp_data.turnover_rate).values.astype("float32"))

    scaler = minMaxScaler(temp_data.net_mf_vol,feature_range=(-1, 1))
    temp_data.net_mf_vol = \
    scaler.transform(pd.DataFrame(temp_data.net_mf_vol).values.astype("float32"))

    scaler = minMaxScaler(temp_data.close)
    temp_data.close = \
    scaler.transform(pd.DataFrame(temp_data.close).values.astype("float32"))
    # #############################################################################
    # #####################################构造数据#################################
    train_data = temp_data[temp_data['trade_date'] <= 20191021]
    test_data = temp_data[temp_data['trade_date'] >= 20190901]

    train_X = pd.DataFrame(train_data).values
    train_Y = pd.DataFrame(train_data['close']).values
    test_X = pd.DataFrame(test_data).values
    test_Y = pd.DataFrame(test_data['close']).values
    len(test_X)

    window_size = 30
    predict_size = 5
    train_X,train_Y = create_dataset(train_X, train_Y, window_size=window_size, predict_size=predict_size)
    test_X,test_Y = create_dataset(test_X, test_Y, window_size=window_size, predict_size=predict_size)
    print(f'train_X.shape: {train_X.shape}   train_Y.shape: {train_Y.shape} \n'
          f'test_X.shape: {test_X.shape}     test_Y.shape: {test_Y.shape} ')
    # #############################################################################
    # #####################################构造模型#################################
    from keras.models import Sequential
    from keras.layers.recurrent import LSTM
    from keras.layers.core import Dense, Activation, Dropout
    from keras.callbacks import EarlyStopping
    early_stopping = EarlyStopping('loss', patience=50)

    input_shape = (train_X.shape[1], train_X.shape[2])
    output_shape = train_Y.shape[1]

    model = Sequential()
    model.add(LSTM(64,input_shape=input_shape, return_sequences=True))
    model.add(Dense(32))
    model.add(LSTM(32))
    model.add(Dense(output_shape))
    model.compile(loss="mean_squared_error",optimizer="adam",metrics=['accuracy'])
    model.summary()
    model.fit(train_X,train_Y,epochs=150,batch_size=32,verbose=2,callbacks=[early_stopping])
    # #############################################################################
    # #####################################预测数值#################################
    train_Y_per = scaler.inverse_transform(model.predict(train_X))
    train_Y_per = pd.DataFrame(train_Y_per)
    train_Y_con = pd.DataFrame(scaler.inverse_transform(train_Y))
    test_Y_per = scaler.inverse_transform(model.predict(test_X))
    test_Y_per = pd.DataFrame(test_Y_per)
    test_Y_con = pd.DataFrame(scaler.inverse_transform(test_Y))
    # #############################################################################
    # #####################################画图检验#################################
    draw_train_Y_per = train_Y_per[-10:].reset_index(drop=True)
    draw_train_Y_con = train_Y_con[-10:].reset_index(drop=True)
    draw_test_Y_per = test_Y_per[-10:].reset_index(drop=True)
    draw_test_Y_con = test_Y_con[-10:].reset_index(drop=True)


    from pyecharts.charts import Line
    from pyecharts import options as opts
    title = '预测数据与实际数据对比图_训练数据'
    path = './picture/line_train.html'
    plt = (
        Line(init_opts=opts.InitOpts(width="1920px", height="900px"))
            .set_global_opts(title_opts=opts.TitleOpts(title=title),
                             yaxis_opts=opts.AxisOpts(),  # 双轴1
                             legend_opts=opts.LegendOpts(type_='scroll',pos_left="left",orient='vertical',pos_top = 'middle'),
                             tooltip_opts=opts.TooltipOpts(trigger='axis'),
                             toolbox_opts=opts.ToolboxOpts(),  # 工具栏
                             datazoom_opts=opts.DataZoomOpts(),  # 缩放功能
							 xaxis_opts=opts.AxisOpts(type_="category") # 行坐标类型
                             )
            .set_series_opts(label_opts=opts.LabelOpts(is_show=False))  # 是否显示数值
            .extend_axis(
            yaxis=opts.AxisOpts()
        )
    )


    for i in range(len(draw_train_Y_per)):
        draw_y1 = [round(i,2) for i in draw_train_Y_per[i:i + 1].T[i].to_list()]
        draw_y2 = [round(i,2) for i in draw_train_Y_con[i:i + 1].T[i].to_list()]

        x_lable = list(range(i,i + draw_train_Y_per.shape[1]))
        temp_plt = (
            Line()
                .add_xaxis(x_lable)
                .add_yaxis('预测数据_' + str(i),draw_y1)
        )
        plt.overlap(temp_plt)
        temp_plt = (
            Line()
                .add_xaxis(x_lable)
                .add_yaxis('实际数据_' + str(i),draw_y2)
        )
        plt.overlap(temp_plt)

    plt.render(path)


    title = '预测数据与实际数据对比图_测试数据'
    path = './picture/line_test.html'
    plt = (
        Line(init_opts=opts.InitOpts(width="1920px", height="900px"))
            .set_global_opts(title_opts=opts.TitleOpts(title=title),
                             yaxis_opts=opts.AxisOpts(),  # 双轴1
                             legend_opts=opts.LegendOpts(type_='scroll', pos_left="left", orient='vertical',
                                                         pos_top='middle'),
                             tooltip_opts=opts.TooltipOpts(trigger='axis'),
                             toolbox_opts=opts.ToolboxOpts(),  # 工具栏
                             datazoom_opts=opts.DataZoomOpts(),  # 缩放功能
							 xaxis_opts=opts.AxisOpts(type_="category") # 行坐标类型
                             )
            .set_series_opts(label_opts=opts.LabelOpts(is_show=False))  # 是否显示数值
            .extend_axis(
            yaxis=opts.AxisOpts()
        )
    )

    for i in range(len(draw_test_Y_per)):
        draw_y1 = [round(i, 2) for i in draw_test_Y_per[i:i + 1].T[i].to_list()]
        draw_y2 = [round(i, 2) for i in draw_test_Y_con[i:i + 1].T[i].to_list()]

        x_lable = list(range(i, i + draw_train_Y_per.shape[1]))
        temp_plt = (
            Line()
                .add_xaxis(x_lable)
                .add_yaxis('预测数据_' + str(i), draw_y1)
        )
        plt.overlap(temp_plt)
        temp_plt = (
            Line()
                .add_xaxis(x_lable)
                .add_yaxis('实际数据_' + str(i), draw_y2)
        )
        plt.overlap(temp_plt)

    plt.render(path)

    # #############################################################################

测试结果
训练数据测试

测试数据测试