复制链接

克隆策略

In [ ]:

In [1]:

# 本代码由可视化策略环境自动生成 2022年9月17日 00:47
# 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 m6_run_bigquant_run(input_1, input_2, input_3):

    df_all = m4.data.read_df()
    df_all = df_all[df_all['date']>'2016-01-01']
    df_all = df_all[(df_all['buy_cond_1']==1)&(df_all['buy_cond_2']==1)&(df_all['buy_cond_3']==1)]
    #df_all['label'] = df_all['y']
    data_1 = DataSource.write_df(df_all)
    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 m6_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 m12_run_bigquant_run(input_1, input_2, input_3):

    df_all = m10.data.read_df()
    df_all = df_all[df_all['date']>'2020-01-01']
    df_all = df_all[(df_all['buy_cond_1']==1)&(df_all['buy_cond_2']==1)&(df_all['buy_cond_3']==1)] 

    data_1 = DataSource.write_df(df_all)
    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 m12_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 m13_run_bigquant_run(input_1, input_2, input_3):
    # 示例代码如下。在这里编写您的代码
    df = pd.DataFrame({'data': [1, 2, 3]})
    data_1 = DataSource.write_df(df)
    data_2 = DataSource.write_pickle(df)
    return Outputs(data_1=data_1, data_2=data_2, 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 m13_post_run_bigquant_run(outputs):
    return outputs

# 回测引擎：初始化函数，只执行一次
def m14_initialize_bigquant_run(context):
    context.set_commission(PerOrder(buy_cost=0.0005, sell_cost=0.0012, min_cost=5))

# 回测引擎：每日数据处理函数，每天执行一次
def m14_handle_data_bigquant_run(context, data):
    # 获取当前持仓
    positions = {e.symbol: p.amount * p.last_sale_price for e, p in context.portfolio.positions.items()}
    
    # 按日期过滤得到今日的预测数据
    df = context.df
    today = data.current_dt.strftime('%Y-%m-%d')
    today_data = df[df['date']==today]
    
    #选择预测值最高的股票，持仓的股票卖出
    df = today_data.sort_values(['pred'],ascending=False)
    df = df[:1]

    buy_instruments = df['instrument'].tolist()
    sell_instruments = [instrument.symbol for instrument in context.portfolio.positions.keys()]
    
    #半仓买入
    cash_for_buy = min(context.portfolio.cash,0.5*context.portfolio.portfolio_value)      
    
    for instrument in sell_instruments:
        context.order_target(context.symbol(instrument), 0)
    for instrument in buy_instruments:
        context.order_value(context.symbol(instrument), cash_for_buy)

# 回测引擎：准备数据，只执行一次
def m14_prepare_bigquant_run(context):
    #模型训练，预测功能也放置在自定义模块
    #-----------------------------------------------------------------------训练---------------------------------------------------------------------------------
    #训练数据，选用五个因子的多因子模型
    train_data = m6.data_1.read_df()
    train_data = train_data[(train_data['date']>'2016-01-01')&(train_data['date']<'2020-01-01')]
    train_data_temp = train_data[['a1', 'a2', 'a3', 'a4', 'a5','y','date','instrument']]
    train_data_temp = train_data_temp.dropna(axis = 0) 

    x = train_data_temp[['a1', 'a2', 'a3', 'a4', 'a5']]
    y = train_data_temp[['y']]
    #用随机森林模型训练
    ##--------------------model---------------------
    from sklearn.ensemble import RandomForestRegressor
    model = RandomForestRegressor(
        n_estimators=10,#弱分类树个数
        bootstrap = False,#不放回抽样

        max_features = None,#考虑所有特征
        max_depth = 5,#最大深度
        min_samples_leaf = 2000#最小样本数
    )
    model.fit(x,y)
    ##--------------------model---------------------
    #------------------------------------------------------------------------训练----------------------------------------------------------------------------------
    
    #---------------------------------------------------------------------------测试------------------------------------------------------------------------------------
    #测试数据
    test_data = m12.data_1.read_df()
    test_data = test_data[(test_data['date']>'2020-01-01')]
  
    test_data_temp = test_data[['a1', 'a2', 'a3', 'a4', 'a5','date','instrument']]
    test_data_temp = test_data_temp.dropna(axis = 0)
    x = test_data_temp[['a1', 'a2', 'a3', 'a4', 'a5']]
    #model预测
    y_pred = model.predict(x)
    
    #整理预测后的数据，预测值要大于1，0.002考虑费用
    df_result = test_data_temp[['date','instrument']]
    df_result['pred'] = y_pred
    df_result = df_result[df_result['pred']>1.002]
    #--------------------------------------------------------------------------测试------------------------------------------------------------------------------------

    context.df = df_result     

# 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 m14_before_trading_start_bigquant_run(context, data):
    pass


m1 = M.instruments.v2(
    start_date='2016-01-01',
    end_date='2020-01-01',
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m2 = M.input_features.v1(
    features="""#收盘突破所有短期均线，收盘表现强势
buy_cond_1 = where((close_0>mean(close_0, 20))&(close_0>mean(close_0, 10))&(close_0>mean(close_0, 5)),1,0)
#底部冲短线均线下方开始反弹
buy_cond_2 = where((low_0<mean(close_0, 5))&(low_0<mean(close_0, 10))&(low_0<mean(close_0, 20)), 1, 0)
#确保当天为阳
buy_cond_3 = where((close_0>open_0), 1, 0)


#预测值Y
y = shift(close_0,-2)/shift(open_0,-1)

# 过滤掉一字涨停
where(shift(high_0, -1) == shift(low_0, -1), NaN, y)

#股价相对位置，RSI指标
a1 = ta_rsi_14_0 
#量价因子
a2 = (ta_ema(volume_0, 12)-ta_ema(volume_0, 26))
#资金流因子，近几天大单相对流通市值
a3 = sum(mf_net_amount_xl_0,10)/market_cap_float_0
#换手率因子
a4 = turn_0/mean(turn_0, 5)
#技术指标因子（量价因子）
a5 = close_0/mean(close_0, 20)"""
)

m3 = M.general_feature_extractor.v7(
    instruments=m1.data,
    features=m2.data,
    start_date='',
    end_date='',
    before_start_days=50
)

m5 = M.chinaa_stock_filter.v1(
    input_data=m3.data,
    index_constituent_cond=['全部'],
    board_cond=['深证主板'],
    industry_cond=['全部'],
    st_cond=['正常'],
    delist_cond=['非退市'],
    output_left_data=False
)

m4 = M.derived_feature_extractor.v3(
    input_data=m5.data,
    features=m2.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=True
)

m6 = M.cached.v3(
    input_1=m4.data,
    run=m6_run_bigquant_run,
    post_run=m6_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m7 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2020-01-01'),
    end_date=T.live_run_param('trading_date', '2022-08-16'),
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m8 = M.input_features.v1(
    features="""#收盘突破所有短期均线，收盘表现强势
buy_cond_1 = where((close_0>mean(close_0, 20))&(close_0>mean(close_0, 10))&(close_0>mean(close_0, 5)),1,0)
#底部冲短线均线下方开始反弹
buy_cond_2 = where((low_0<mean(close_0, 5))&(low_0<mean(close_0, 10))&(low_0<mean(close_0, 20)), 1, 0)
#确保当天为阳
buy_cond_3 = where((close_0>open_0), 1, 0)


#预测值Y
y = shift(close_0,-2)/shift(open_0,-1)

# 过滤掉一字涨停
where(shift(high_0, -1) == shift(low_0, -1), NaN, y)

#股价相对位置，RSI指标
a1 = ta_rsi_14_0 
#量价因子
a2 = (ta_ema(volume_0, 12)-ta_ema(volume_0, 26))
#资金流因子，近几天大单相对流通市值
a3 = sum(mf_net_amount_xl_0,10)/market_cap_float_0
#换手率因子
a4 = turn_0/mean(turn_0, 5)
#技术指标因子（量价因子）
a5 = close_0/mean(close_0, 20)"""
)

m9 = M.general_feature_extractor.v7(
    instruments=m7.data,
    features=m8.data,
    start_date='',
    end_date='',
    before_start_days=50
)

m11 = M.chinaa_stock_filter.v1(
    input_data=m9.data,
    index_constituent_cond=['全部'],
    board_cond=['深证主板'],
    industry_cond=['全部'],
    st_cond=['正常'],
    delist_cond=['非退市'],
    output_left_data=False
)

m10 = M.derived_feature_extractor.v3(
    input_data=m11.data,
    features=m8.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=True
)

m12 = M.cached.v3(
    input_1=m10.data,
    run=m12_run_bigquant_run,
    post_run=m12_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m13 = M.cached.v3(
    input_1=m6.data_1,
    input_2=m12.data_1,
    run=m13_run_bigquant_run,
    post_run=m13_post_run_bigquant_run,
    input_ports='',
    params='{}',
    output_ports=''
)

m14 = M.trade.v4(
    instruments=m7.data,
    options_data=m13.data_1,
    start_date='',
    end_date='',
    initialize=m14_initialize_bigquant_run,
    handle_data=m14_handle_data_bigquant_run,
    prepare=m14_prepare_bigquant_run,
    before_trading_start=m14_before_trading_start_bigquant_run,
    volume_limit=0.025,
    order_price_field_buy='open',
    order_price_field_sell='close',
    capital_base=100000,
    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.HIX'
)

[2022-09-17 00:43:55.336150] INFO: moduleinvoker: instruments.v2 开始运行..

[2022-09-17 00:43:55.354087] INFO: moduleinvoker: 命中缓存

[2022-09-17 00:43:55.355950] INFO: moduleinvoker: instruments.v2 运行完成[0.019819s].

[2022-09-17 00:43:55.364067] INFO: moduleinvoker: input_features.v1 开始运行..

[2022-09-17 00:43:55.404786] INFO: moduleinvoker: input_features.v1 运行完成[0.040706s].

[2022-09-17 00:43:55.427818] INFO: moduleinvoker: general_feature_extractor.v7 开始运行..

[2022-09-17 00:43:55.445625] INFO: moduleinvoker: 命中缓存

[2022-09-17 00:43:55.449097] INFO: moduleinvoker: general_feature_extractor.v7 运行完成[0.021284s].

[2022-09-17 00:43:55.474911] INFO: moduleinvoker: chinaa_stock_filter.v1 开始运行..

[2022-09-17 00:43:55.489105] INFO: moduleinvoker: 命中缓存

[2022-09-17 00:43:55.491487] INFO: moduleinvoker: chinaa_stock_filter.v1 运行完成[0.016578s].

[2022-09-17 00:43:55.507609] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2022-09-17 00:44:00.293390] INFO: derived_feature_extractor: 提取完成 buy_cond_1 = where((close_0>mean(close_0, 20))&(close_0>mean(close_0, 10))&(close_0>mean(close_0, 5)),1,0), 2.299s

[2022-09-17 00:44:02.502300] INFO: derived_feature_extractor: 提取完成 buy_cond_2 = where((low_0[2022-09-17 00:44:02.508384] INFO: derived_feature_extractor: 提取完成 buy_cond_3 = where((close_0>open_0), 1, 0), 0.004s
[2022-09-17 00:44:02.874753] INFO: derived_feature_extractor: 提取完成 y = shift(close_0,-2)/shift(open_0,-1), 0.365s
[2022-09-17 00:44:03.213000] INFO: derived_feature_extractor: 提取完成 where(shift(high_0, -1) == shift(low_0, -1), NaN, y), 0.337s
[2022-09-17 00:44:03.216662] INFO: derived_feature_extractor: 提取完成 a1 = ta_rsi_14_0, 0.002s
[2022-09-17 00:44:07.234698] INFO: derived_feature_extractor: 提取完成 a2 = (ta_ema(volume_0, 12)-ta_ema(volume_0, 26)), 4.017s
[2022-09-17 00:44:07.978205] INFO: derived_feature_extractor: 提取完成 a3 = sum(mf_net_amount_xl_0,10)/market_cap_float_0, 0.742s
[2022-09-17 00:44:08.716234] INFO: derived_feature_extractor: 提取完成 a4 = turn_0/mean(turn_0, 5), 0.736s
[2022-09-17 00:44:09.442628] INFO: derived_feature_extractor: 提取完成 a5 = close_0/mean(close_0, 20), 0.725s
[2022-09-17 00:44:09.731956] INFO: derived_feature_extractor: /y_2015, 38199
[2022-09-17 00:44:10.175959] INFO: derived_feature_extractor: /y_2016, 270989
[2022-09-17 00:44:10.802658] INFO: derived_feature_extractor: /y_2017, 292568
[2022-09-17 00:44:11.435119] INFO: derived_feature_extractor: /y_2018, 308078
[2022-09-17 00:44:12.068629] INFO: derived_feature_extractor: /y_2019, 323494
[2022-09-17 00:44:12.417148] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[16.909531s].
[2022-09-17 00:44:12.432985] INFO: moduleinvoker: cached.v3 开始运行..
[2022-09-17 00:44:13.605393] INFO: moduleinvoker: cached.v3 运行完成[1.172412s].
[2022-09-17 00:44:14.036639] INFO: moduleinvoker: instruments.v2 开始运行..
[2022-09-17 00:44:14.048326] INFO: moduleinvoker: 命中缓存
[2022-09-17 00:44:14.050383] INFO: moduleinvoker: instruments.v2 运行完成[0.013755s].
[2022-09-17 00:44:14.056240] INFO: moduleinvoker: input_features.v1 开始运行..
[2022-09-17 00:44:14.063441] INFO: moduleinvoker: 命中缓存
[2022-09-17 00:44:14.065266] INFO: moduleinvoker: input_features.v1 运行完成[0.009041s].
[2022-09-17 00:44:14.095543] INFO: moduleinvoker: general_feature_extractor.v7 开始运行..
[2022-09-17 00:44:20.771528] INFO: 基础特征抽取: 年份 2019, 特征行数=133862
[2022-09-17 00:44:29.261380] INFO: 基础特征抽取: 年份 2020, 特征行数=945961
[2022-09-17 00:44:36.834478] INFO: 基础特征抽取: 年份 2021, 特征行数=1061527
[2022-09-17 00:44:42.360621] INFO: 基础特征抽取: 年份 2022, 特征行数=714682
[2022-09-17 00:44:42.506627] INFO: 基础特征抽取: 总行数: 2856032
[2022-09-17 00:44:42.514095] INFO: moduleinvoker: general_feature_extractor.v7 运行完成[28.418575s].
[2022-09-17 00:44:42.535315] INFO: moduleinvoker: chinaa_stock_filter.v1 开始运行..
[2022-09-17 00:44:43.549007] INFO: A股股票过滤: 过滤 /y_2019, 47663/0/133862
[2022-09-17 00:44:48.268367] INFO: A股股票过滤: 过滤 /y_2020, 320025/0/945961
[2022-09-17 00:44:53.460273] INFO: A股股票过滤: 过滤 /y_2021, 333350/0/1061527
[2022-09-17 00:44:57.147987] INFO: A股股票过滤: 过滤 /y_2022, 209965/0/714682
[2022-09-17 00:44:57.153352] INFO: A股股票过滤: 过滤完成, 911003 + 0
[2022-09-17 00:44:57.188238] INFO: moduleinvoker: chinaa_stock_filter.v1 运行完成[14.652922s].
[2022-09-17 00:44:57.196914] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..
[2022-09-17 00:45:00.623625] INFO: derived_feature_extractor: 提取完成 buy_cond_1 = where((close_0>mean(close_0, 20))&(close_0>mean(close_0, 10))&(close_0>mean(close_0, 5)),1,0), 1.795s
[2022-09-17 00:45:02.316919] INFO: derived_feature_extractor: 提取完成 buy_cond_2 = where((low_0[2022-09-17 00:45:02.321585] INFO: derived_feature_extractor: 提取完成 buy_cond_3 = where((close_0>open_0), 1, 0), 0.003s
[2022-09-17 00:45:02.624873] INFO: derived_feature_extractor: 提取完成 y = shift(close_0,-2)/shift(open_0,-1), 0.302s
[2022-09-17 00:45:02.912175] INFO: derived_feature_extractor: 提取完成 where(shift(high_0, -1) == shift(low_0, -1), NaN, y), 0.285s
[2022-09-17 00:45:02.916522] INFO: derived_feature_extractor: 提取完成 a1 = ta_rsi_14_0, 0.002s
[2022-09-17 00:45:06.810362] INFO: derived_feature_extractor: 提取完成 a2 = (ta_ema(volume_0, 12)-ta_ema(volume_0, 26)), 3.892s
[2022-09-17 00:45:07.432014] INFO: derived_feature_extractor: 提取完成 a3 = sum(mf_net_amount_xl_0,10)/market_cap_float_0, 0.620s
[2022-09-17 00:45:08.030647] INFO: derived_feature_extractor: 提取完成 a4 = turn_0/mean(turn_0, 5), 0.597s
[2022-09-17 00:45:08.626764] INFO: derived_feature_extractor: 提取完成 a5 = close_0/mean(close_0, 20), 0.594s
[2022-09-17 00:45:08.837870] INFO: derived_feature_extractor: /y_2019, 47663
[2022-09-17 00:45:09.425888] INFO: derived_feature_extractor: /y_2020, 320025
[2022-09-17 00:45:10.538496] INFO: derived_feature_extractor: /y_2021, 333350
[2022-09-17 00:45:11.035476] INFO: derived_feature_extractor: /y_2022, 209965
[2022-09-17 00:45:11.279335] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[14.0824s].
[2022-09-17 00:45:11.306928] INFO: moduleinvoker: cached.v3 开始运行..
[2022-09-17 00:45:12.099611] INFO: moduleinvoker: cached.v3 运行完成[0.792684s].
[2022-09-17 00:45:12.113225] INFO: moduleinvoker: cached.v3 开始运行..
[2022-09-17 00:45:12.187563] INFO: moduleinvoker: cached.v3 运行完成[0.074345s].
[2022-09-17 00:45:16.061046] INFO: moduleinvoker: backtest.v8 开始运行..
[2022-09-17 00:45:16.067148] INFO: backtest: biglearning backtest:V8.6.3
[2022-09-17 00:45:18.018423] INFO: backtest: product_type:stock by specified
[2022-09-17 00:45:18.360733] INFO: moduleinvoker: cached.v2 开始运行..
[2022-09-17 00:45:39.202748] INFO: backtest: 读取股票行情完成:4050916
[2022-09-17 00:45:42.691513] INFO: moduleinvoker: cached.v2 运行完成[24.330785s].
[2022-09-17 00:45:46.804697] INFO: backtest: algo history_data=DataSource(c1763b1b67d745dc8378a7077307f102T)
[2022-09-17 00:45:46.806561] INFO: algo: TradingAlgorithm V1.8.8
[2022-09-17 00:45:48.467341] INFO: algo: trading transform...
[2022-09-17 00:45:50.250345] INFO: algo: handle_splits get splits [dt:2020-07-03 00:00:00+00:00] [asset:Equity(1701 [002973.SZA]), ratio:0.9987605810165405]
[2022-09-17 00:45:50.855161] INFO: algo: handle_splits get splits [dt:2020-10-19 00:00:00+00:00] [asset:Equity(4431 [002088.SZA]), ratio:0.9682971239089966]
[2022-09-17 00:45:52.236317] INFO: algo: handle_splits get splits [dt:2021-07-05 00:00:00+00:00] [asset:Equity(574 [000671.SZA]), ratio:0.9278937578201294]
[2022-09-17 00:45:52.238011] INFO: Position: position stock handle split[sid:574, orig_amount:16400, new_amount:17674.0, orig_cost:5.290000539518855, new_cost:4.9086, ratio:0.9278937578201294, last_sale_price:4.889999866485596]
[2022-09-17 00:45:52.239356] INFO: Position: after split: PositionStock(asset:Equity(574 [000671.SZA]), amount:17674.0, cost_basis:4.9086, last_sale_price:5.269999980926514)
[2022-09-17 00:45:52.240989] INFO: Position: returning cash: 2.1382
[2022-09-17 00:45:54.031696] INFO: algo: handle_splits get splits [dt:2022-06-16 00:00:00+00:00] [asset:Equity(1394 [000069.SZA]), ratio:0.982332170009613]
[2022-09-17 00:45:54.374831] INFO: Performance: Simulated 636 trading days out of 636.
[2022-09-17 00:45:54.376365] INFO: Performance: first open: 2020-01-02 09:30:00+00:00
[2022-09-17 00:45:54.377516] INFO: Performance: last close: 2022-08-16 15:00:00+00:00
[2022-09-17 00:46:00.579695] INFO: moduleinvoker: backtest.v8 运行完成[44.518652s].
[2022-09-17 00:46:00.581477] INFO: moduleinvoker: trade.v4 运行完成[48.37322s].

收益率276.78%
年化收益率69.15%
基准收益率1.98%
阿尔法0.76
贝塔0.64
夏普比率1.53
胜率0.51
盈亏比1.41
收益波动率36.75%
信息比率0.1
最大回撤25.01%

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-799cdee88aa24e259999adbd784ff144"}/bigcharts-data-end