克隆策略

使用深度学习技术预测股票价格¶

In [22]:

# 本代码由可视化策略环境自动生成 2020年2月28日 09:19
# 本代码单元只能在可视化模式下编辑。您也可以拷贝代码，粘贴到新建的代码单元或者策略，然后修改。


# Python 代码入口函数，input_1/2/3 对应三个输入端，data_1/2/3 对应三个输出端
def m24_run_bigquant_run(input_1, input_2, input_3):
    fields = ['open','high','low','close','volume']
    input_1_df = input_1.read_pickle()
    ins = input_1_df['instruments']
    start_date = input_1_df['start_date']
    end_date = input_1_df['end_date']
    df = D.history_data(ins, start_date, end_date, fields)     
    data_1 = DataSource.write_df(df)
    return Outputs(data_1=data_1, data_2=None, data_3=None)

# 后处理函数，可选。输入是主函数的输出，可以在这里对数据做处理，或者返回更友好的outputs数据格式。此函数输出不会被缓存。
def m24_post_run_bigquant_run(outputs):
    return outputs

# Python 代码入口函数，input_1/2/3 对应三个输入端，data_1/2/3 对应三个输出端
def m16_run_bigquant_run(input_1, input_2, input_3):
    input_ds = input_1
    df = input_ds.read_df()
    df['return'] = (df.close.shift(-10)/df.close - 1)
    df['label'] = np.where(df['return'] > 0, 1, 0)
    ds = DataSource.write_df(df)
    return Outputs(data_1=ds)


# 后处理函数，可选。输入是主函数的输出，可以在这里对数据做处理，或者返回更友好的outputs数据格式。此函数输出不会被缓存。
def m16_post_run_bigquant_run(outputs):
    return outputs

# Python 代码入口函数，input_1/2/3 对应三个输入端，data_1/2/3 对应三个输出端
def m2_run_bigquant_run(input_1, input_2, input_3):
    input_series = input_1
    input_df = input_2
    test_data = input_df.read_pickle()
    pred_label = input_series.read_pickle()
  
    pred_result = pred_label.reshape(pred_label.shape[0]) 
    dt = input_3.read_df()['date'][-1*len(pred_result):]
    pred_df = pd.Series(pred_result, index=dt)
    ds = DataSource.write_df(pred_df)
    
    pred_label = np.where(pred_label>0.5,1,0)
    labels = test_data['y']
    print('准确率%s'%(np.mean(pred_label==labels)))
    
    return Outputs(data_1=ds)

# 后处理函数，可选。输入是主函数的输出，可以在这里对数据做处理，或者返回更友好的outputs数据格式。此函数输出不会被缓存。
def m2_post_run_bigquant_run(outputs):
    return outputs

# 回测引擎：初始化函数，只执行一次
def m29_initialize_bigquant_run(context):
    # 加载预测数据
    context.prediction = context.options['data'].read_df()

    # 系统已经设置了默认的交易手续费和滑点，要修改手续费可使用如下函数
    context.set_commission(PerOrder(buy_cost=0.0003, sell_cost=0.0013, min_cost=5))
# 回测引擎：每日数据处理函数，每天执行一次
def m29_handle_data_bigquant_run(context, data):
    # 按日期过滤得到今日的预测数据
    try:
        prediction = context.prediction[data.current_dt.strftime('%Y-%m-%d')]
    except KeyError as e:
        return

   
    
    instrument = context.instruments[0]
    sid = context.symbol(instrument)
    cur_position = context.portfolio.positions[sid].amount
    #print('date: ',data.current_dt, '持仓： ', cur_position)
    
    # 交易逻辑
    if prediction > 0.5 and cur_position == 0:
        context.order_target_percent(context.symbol(instrument), 1)
        print(data.current_dt, '买入！')
        
    elif prediction < 0.5 and cur_position > 0:
        context.order_target_percent(context.symbol(instrument), 0)
        print(data.current_dt, '卖出！')
    
# 回测引擎：准备数据，只执行一次
def m29_prepare_bigquant_run(context):
    pass

# 回测引擎：每个单位时间开始前调用一次，即每日开盘前调用一次。
def m29_before_trading_start_bigquant_run(context, data):
    pass


m3 = M.dl_layer_input.v1(
    shape='50,5',
    batch_shape='',
    dtype='float32',
    sparse=False,
    name=''
)

m13 = M.dl_layer_reshape.v1(
    inputs=m3.data,
    target_shape='50,5,1',
    name=''
)

m14 = M.dl_layer_conv2d.v1(
    inputs=m13.data,
    filters=32,
    kernel_size='3,5',
    strides='1,1',
    padding='valid',
    data_format='channels_last',
    dilation_rate='1,1',
    activation='relu',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    bias_initializer='Zeros',
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    bias_constraint='None',
    name=''
)

m15 = M.dl_layer_reshape.v1(
    inputs=m14.data,
    target_shape='48,32',
    name=''
)

m4 = M.dl_layer_lstm.v1(
    inputs=m15.data,
    units=32,
    activation='tanh',
    recurrent_activation='hard_sigmoid',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    recurrent_initializer='Orthogonal',
    bias_initializer='Ones',
    unit_forget_bias=True,
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    recurrent_regularizer='None',
    recurrent_regularizer_l1=0,
    recurrent_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    recurrent_constraint='None',
    bias_constraint='None',
    dropout=0,
    recurrent_dropout=0,
    return_sequences=False,
    implementation='1',
    name=''
)

m11 = M.dl_layer_dropout.v1(
    inputs=m4.data,
    rate=0.2,
    noise_shape='',
    name=''
)

m10 = M.dl_layer_dense.v1(
    inputs=m11.data,
    units=32,
    activation='tanh',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    bias_initializer='Zeros',
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    bias_constraint='None',
    name=''
)

m12 = M.dl_layer_dropout.v1(
    inputs=m10.data,
    rate=0.2,
    noise_shape='',
    name=''
)

m9 = M.dl_layer_dense.v1(
    inputs=m12.data,
    units=1,
    activation='sigmoid',
    use_bias=True,
    kernel_initializer='glorot_uniform',
    bias_initializer='Zeros',
    kernel_regularizer='None',
    kernel_regularizer_l1=0,
    kernel_regularizer_l2=0,
    bias_regularizer='None',
    bias_regularizer_l1=0,
    bias_regularizer_l2=0,
    activity_regularizer='None',
    activity_regularizer_l1=0,
    activity_regularizer_l2=0,
    kernel_constraint='None',
    bias_constraint='None',
    name=''
)

m5 = M.dl_model_init.v1(
    inputs=m3.data,
    outputs=m9.data
)

m8 = M.input_features.v1(
    features="""(close/shift(close,1)-1)*10
(high/shift(high,1)-1)*10
(low/shift(low,1)-1)*10
(open/shift(open,1)-1)*10
(volume/shift(volume,1)-1)*10"""
)

m25 = M.instruments.v2(
    start_date='2015-01-01',
    end_date='2018-02-07',
    market='CN_STOCK_A',
    instrument_list='600009.SHA',
    max_count=0
)

m24 = M.cached.v3(
    input_1=m25.data,
    run=m24_run_bigquant_run,
    post_run=m24_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m16 = M.cached.v3(
    input_1=m24.data_1,
    run=m16_run_bigquant_run,
    post_run=m16_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m17 = M.derived_feature_extractor.v3(
    input_data=m24.data_1,
    features=m8.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False,
    user_functions={}
)

m18 = M.join.v3(
    data1=m16.data_1,
    data2=m17.data,
    on='date,instrument',
    how='inner',
    sort=True
)

m19 = M.dropnan.v1(
    input_data=m18.data
)

m20 = M.filter.v3(
    input_data=m19.data,
    expr='date<\'2017-03-01\'',
    output_left_data=False
)

m27 = M.dl_convert_to_bin.v2(
    input_data=m20.data,
    features=m8.data,
    window_size=50,
    feature_clip=5,
    flatten=False,
    window_along_col='instrument'
)

m6 = M.dl_model_train.v1(
    input_model=m5.data,
    training_data=m27.data,
    optimizer='Adam',
    loss='binary_crossentropy',
    metrics='accuracy',
    batch_size=2048,
    epochs=10,
    n_gpus=1,
    verbose='1:输出进度条记录'
)

m21 = M.filter.v3(
    input_data=m19.data,
    expr='date>\'2017-03-01\'',
    output_left_data=False
)

m28 = M.dl_convert_to_bin.v2(
    input_data=m21.data,
    features=m8.data,
    window_size=50,
    feature_clip=5,
    flatten=False,
    window_along_col='instrument'
)

m7 = M.dl_model_predict.v1(
    trained_model=m6.data,
    input_data=m28.data,
    batch_size=10240,
    n_gpus=2,
    verbose='2:每个epoch输出一行记录'
)

m2 = M.cached.v3(
    input_1=m7.data,
    input_2=m28.data,
    input_3=m21.data,
    run=m2_run_bigquant_run,
    post_run=m2_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m29 = M.trade.v4(
    instruments=m25.data,
    options_data=m2.data_1,
    start_date='2017-04-01',
    end_date='',
    initialize=m29_initialize_bigquant_run,
    handle_data=m29_handle_data_bigquant_run,
    prepare=m29_prepare_bigquant_run,
    before_trading_start=m29_before_trading_start_bigquant_run,
    volume_limit=0.025,
    order_price_field_buy='open',
    order_price_field_sell='close',
    capital_base=1000000,
    auto_cancel_non_tradable_orders=True,
    data_frequency='daily',
    price_type='真实价格',
    product_type='股票',
    plot_charts=True,
    backtest_only=False,
    benchmark='000300.SHA'
)

[2020-02-28 09:18:54.528724] INFO: bigquant: dl_layer_input.v1 运行完成[0.003536s].

[2020-02-28 09:18:54.547210] INFO: bigquant: dl_layer_reshape.v1 运行完成[0.0154s].

[2020-02-28 09:18:54.574590] INFO: bigquant: dl_layer_conv2d.v1 运行完成[0.0206s].

[2020-02-28 09:18:54.591438] INFO: bigquant: dl_layer_reshape.v1 运行完成[0.013799s].

[2020-02-28 09:18:54.910879] INFO: bigquant: dl_layer_lstm.v1 运行完成[0.316033s].

[2020-02-28 09:18:54.948990] INFO: bigquant: dl_layer_dropout.v1 运行完成[0.035651s].

[2020-02-28 09:18:54.973678] INFO: bigquant: dl_layer_dense.v1 运行完成[0.020963s].

[2020-02-28 09:18:55.018405] INFO: bigquant: dl_layer_dropout.v1 运行完成[0.042228s].

[2020-02-28 09:18:55.040589] INFO: bigquant: dl_layer_dense.v1 运行完成[0.018656s].

[2020-02-28 09:18:55.069337] INFO: bigquant: cached.v3 开始运行..

[2020-02-28 09:18:55.186916] INFO: bigquant: cached.v3 运行完成[0.11758s].

[2020-02-28 09:18:55.188841] INFO: bigquant: dl_model_init.v1 运行完成[0.145496s].

[2020-02-28 09:18:55.191837] INFO: bigquant: input_features.v1 开始运行..

[2020-02-28 09:18:55.240015] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.242651] INFO: bigquant: input_features.v1 运行完成[0.050786s].

[2020-02-28 09:18:55.245912] INFO: bigquant: instruments.v2 开始运行..

[2020-02-28 09:18:55.304140] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.305916] INFO: bigquant: instruments.v2 运行完成[0.059989s].

[2020-02-28 09:18:55.310426] INFO: bigquant: cached.v3 开始运行..

[2020-02-28 09:18:55.383621] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.385383] INFO: bigquant: cached.v3 运行完成[0.074954s].

[2020-02-28 09:18:55.388902] INFO: bigquant: cached.v3 开始运行..

[2020-02-28 09:18:55.464518] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.466362] INFO: bigquant: cached.v3 运行完成[0.077448s].

[2020-02-28 09:18:55.468879] INFO: bigquant: derived_feature_extractor.v3 开始运行..

[2020-02-28 09:18:55.509595] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.511694] INFO: bigquant: derived_feature_extractor.v3 运行完成[0.042803s].

[2020-02-28 09:18:55.514626] INFO: bigquant: join.v3 开始运行..

[2020-02-28 09:18:55.561108] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.563634] INFO: bigquant: join.v3 运行完成[0.048985s].

[2020-02-28 09:18:55.567879] INFO: bigquant: dropnan.v1 开始运行..

[2020-02-28 09:18:55.611799] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.613558] INFO: bigquant: dropnan.v1 运行完成[0.045681s].

[2020-02-28 09:18:55.616112] INFO: bigquant: filter.v3 开始运行..

[2020-02-28 09:18:55.662425] INFO: bigquant: 命中缓存

[2020-02-28 09:18:55.664379] INFO: bigquant: filter.v3 运行完成[0.048262s].

[2020-02-28 09:18:55.705850] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2020-02-28 09:18:56.171083] INFO: bigquant: dl_convert_to_bin.v2 运行完成[0.465224s].

[2020-02-28 09:18:56.174539] INFO: bigquant: dl_model_train.v1 开始运行..

[2020-02-28 09:18:57.592991] INFO: dl_model_train: 准备训练，训练样本个数：523，迭代次数：10

[2020-02-28 09:19:12.132575] INFO: dl_model_train: 训练结束，耗时：14.54s

[2020-02-28 09:19:12.352700] INFO: bigquant: dl_model_train.v1 运行完成[16.178163s].

[2020-02-28 09:19:12.356029] INFO: bigquant: filter.v3 开始运行..

[2020-02-28 09:19:12.471868] INFO: filter: 使用表达式 date>'2017-03-01' 过滤

[2020-02-28 09:19:12.586971] INFO: filter: 过滤 /data, 224/0/748

[2020-02-28 09:19:12.684508] INFO: bigquant: filter.v3 运行完成[0.328462s].

[2020-02-28 09:19:12.722692] INFO: bigquant: dl_convert_to_bin.v2 开始运行..

[2020-02-28 09:19:13.016660] INFO: bigquant: dl_convert_to_bin.v2 运行完成[0.29397s].

[2020-02-28 09:19:13.025154] INFO: bigquant: dl_model_predict.v1 开始运行..

[2020-02-28 09:19:14.327365] INFO: device_manager: 没有gpu资源，将使用cpu计算

[2020-02-28 09:19:14.332116] INFO: device_manager: 本次操作不使用GPU

[2020-02-28 09:19:15.015724] INFO: bigquant: dl_model_predict.v1 运行完成[1.990562s].

[2020-02-28 09:19:15.023041] INFO: bigquant: cached.v3 开始运行..

[2020-02-28 09:19:15.683373] INFO: bigquant: cached.v3 运行完成[0.66032s].

[2020-02-28 09:19:17.467537] INFO: bigquant: backtest.v8 开始运行..

[2020-02-28 09:19:17.471485] INFO: bigquant: biglearning backtest:V8.3.2

[2020-02-28 09:19:17.473040] INFO: bigquant: product_type:stock by specified

[2020-02-28 09:19:17.777709] INFO: bigquant: cached.v2 开始运行..

[2020-02-28 09:19:28.389830] INFO: bigquant: 读取股票行情完成:456

[2020-02-28 09:19:29.220944] INFO: bigquant: cached.v2 运行完成[11.443253s].

[2020-02-28 09:19:29.279800] INFO: algo: TradingAlgorithm V1.6.5

[2020-02-28 09:19:29.356952] INFO: algo: trading transform...

[2020-02-28 09:19:29.609371] INFO: algo: handle_splits get splits [dt:2017-08-24 00:00:00+00:00] [asset:Equity(0 [600009.SHA]), ratio:0.9882099126332605]

[2020-02-28 09:19:29.611292] INFO: Position: position stock handle split[sid:0, orig_amount:33000, new_amount:33393.0, orig_cost:30.280028020664357, new_cost:29.92, ratio:0.9882099126332605, last_sale_price:36.87999363789555]

[2020-02-28 09:19:29.612533] INFO: Position: after split: PositionStock(asset:Equity(0 [600009.SHA]), amount:33393.0, cost_basis:29.92, last_sale_price:37.31999969482422)

[2020-02-28 09:19:29.618782] INFO: Position: returning cash: 26.36

[2020-02-28 09:19:29.906879] INFO: Performance: Simulated 212 trading days out of 212.

[2020-02-28 09:19:29.908462] INFO: Performance: first open: 2017-04-05 09:30:00+00:00

[2020-02-28 09:19:29.910045] INFO: Performance: last close: 2018-02-07 15:00:00+00:00

[2020-02-28 09:19:33.340547] INFO: bigquant: backtest.v8 运行完成[15.872986s].

[2020-02-28 09:19:33.343459] INFO: bigquant: trade.v4 运行完成[17.639216s].

Train on 523 samples
Epoch 1/10
523/523 [==============================] - 8s 15ms/sample - loss: 0.7726 - accuracy: 0.4551
Epoch 2/10
523/523 [==============================] - 1s 2ms/sample - loss: 0.7529 - accuracy: 0.4589
Epoch 3/10
523/523 [==============================] - 1s 2ms/sample - loss: 0.7388 - accuracy: 0.5067
Epoch 4/10
523/523 [==============================] - 1s 1ms/sample - loss: 0.7166 - accuracy: 0.4780
Epoch 5/10
523/523 [==============================] - 1s 2ms/sample - loss: 0.7286 - accuracy: 0.5182
Epoch 6/10
523/523 [==============================] - 1s 2ms/sample - loss: 0.7243 - accuracy: 0.5143
Epoch 7/10
523/523 [==============================] - 1s 1ms/sample - loss: 0.7026 - accuracy: 0.5564
Epoch 8/10
523/523 [==============================] - 1s 1ms/sample - loss: 0.7166 - accuracy: 0.5488
Epoch 9/10
523/523 [==============================] - 1s 1ms/sample - loss: 0.7139 - accuracy: 0.5698
Epoch 10/10
523/523 [==============================] - 1s 1ms/sample - loss: 0.7125 - accuracy: 0.5621

224/224 - 0s
DataSource(2721d381a86742458ec4a4140c9aad11T, v3)

准确率0.7053571428571429

2017-04-05 15:00:00+00:00 买入！

收益率59.37%
年化收益率74.01%
基准收益率17.2%
阿尔法0.46
贝塔0.69
夏普比率1.86
胜率1.0
盈亏比0.0
收益波动率30.65%
信息比率0.09
最大回撤11.53%

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-0cbb7c0c5c1143b0aa993794889433df"}/bigcharts-data-end