克隆策略

StockRanker多因子选股策略¶

In [2]:

# 本代码由可视化策略环境自动生成 2021年12月1日 14:47
# 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。


# 回测引擎：初始化函数，只执行一次
def m4_initialize_bigquant_run(context):
    # 加载预测数据
    context.ranker_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))
    # 预测数据，通过options传入进来，使用 read_df 函数，加载到内存 (DataFrame)
    # 设置买入的股票数量，这里买入预测股票列表排名靠前的5只
    stock_count = 5
    # 每只的股票的权重，如下的权重分配会使得靠前的股票分配多一点的资金，[0.339160, 0.213986, 0.169580, ..]
    context.stock_weights = T.norm([1 / math.log(i + 2) for i in range(0, stock_count)])
    # 设置每只股票占用的最大资金比例
    context.max_cash_per_instrument = 0.2
    context.options['hold_days'] = 5

# 回测引擎：每日数据处理函数，每天执行一次
def m4_handle_data_bigquant_run(context, data):
    # 按日期过滤得到今日的预测数据
    ranker_prediction = context.ranker_prediction[
        context.ranker_prediction.date == data.current_dt.strftime('%Y-%m-%d')]

    # 1. 资金分配
    # 平均持仓时间是hold_days，每日都将买入股票，每日预期使用 1/hold_days 的资金
    # 实际操作中，会存在一定的买入误差，所以在前hold_days天，等量使用资金；之后，尽量使用剩余资金（这里设置最多用等量的1.5倍）
    is_staging = context.trading_day_index < context.options['hold_days'] # 是否在建仓期间（前 hold_days 天）
    cash_avg = context.portfolio.portfolio_value / context.options['hold_days']
    cash_for_buy = min(context.portfolio.cash, (1 if is_staging else 1.5) * cash_avg)
    cash_for_sell = cash_avg - (context.portfolio.cash - cash_for_buy)
    positions = {e.symbol: p.amount * p.last_sale_price
                 for e, p in context.perf_tracker.position_tracker.positions.items()}

    # 2. 生成卖出订单：hold_days天之后才开始卖出；对持仓的股票，按机器学习算法预测的排序末位淘汰
    if not is_staging and cash_for_sell > 0:
        equities = {e.symbol: e for e, p in context.perf_tracker.position_tracker.positions.items()}
        instruments = list(reversed(list(ranker_prediction.instrument[ranker_prediction.instrument.apply(
                lambda x: x in equities and not context.has_unfinished_sell_order(equities[x]))])))
        # print('rank order for sell %s' % instruments)
        for instrument in instruments:
            context.order_target(context.symbol(instrument), 0)
            cash_for_sell -= positions[instrument]
            if cash_for_sell <= 0:
                break

    # 3. 生成买入订单：按机器学习算法预测的排序，买入前面的stock_count只股票
    buy_cash_weights = context.stock_weights
    buy_instruments = list(ranker_prediction.instrument[:len(buy_cash_weights)])
    max_cash_per_instrument = context.portfolio.portfolio_value * context.max_cash_per_instrument
    for i, instrument in enumerate(buy_instruments):
        cash = cash_for_buy * buy_cash_weights[i]
        if cash > max_cash_per_instrument - positions.get(instrument, 0):
            # 确保股票持仓量不会超过每次股票最大的占用资金量
            cash = max_cash_per_instrument - positions.get(instrument, 0)
        if cash > 0:
            context.order_value(context.symbol(instrument), cash)

# 回测引擎：准备数据，只执行一次
def m4_prepare_bigquant_run(context):
    pass


m1 = M.instruments.v2(
    start_date='2010-01-01',
    end_date='2015-01-01',
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m2 = M.advanced_auto_labeler.v2(
    instruments=m1.data,
    label_expr="""# #号开始的表示注释
# 0. 每行一个，顺序执行，从第二个开始，可以使用label字段
# 1. 可用数据字段见 https://bigquant.com/docs/data_history_data.html
#   添加benchmark_前缀，可使用对应的benchmark数据
# 2. 可用操作符和函数见 `表达式引擎 <https://bigquant.com/docs/big_expr.html>`_

# 计算收益：5日收盘价(作为卖出价格)除以明日开盘价(作为买入价格)
shift(close, -5) / shift(open, -1)

# 极值处理：用1%和99%分位的值做clip
clip(label, all_quantile(label, 0.01), all_quantile(label, 0.99))

# 将分数映射到分类，这里使用20个分类
all_wbins(label, 20)

# 过滤掉一字涨停的情况 (设置label为NaN，在后续处理和训练中会忽略NaN的label)
where(shift(high, -1) == shift(low, -1), NaN, label)
""",
    start_date='',
    end_date='',
    benchmark='000300.SHA',
    drop_na_label=True,
    cast_label_int=True
)

m3 = M.input_features.v1(
    features="""# #号开始的表示注释
# 多个特征，每行一个，可以包含基础特征和衍生特征
return_5
return_10
return_20
avg_amount_0/avg_amount_5
avg_amount_5/avg_amount_20
rank_avg_amount_0/rank_avg_amount_5
rank_avg_amount_5/rank_avg_amount_10
rank_return_0
rank_return_5
rank_return_10
rank_return_0/rank_return_5
rank_return_5/rank_return_10
pe_ttm_0
"""
)

m15 = M.general_feature_extractor.v7(
    instruments=m1.data,
    features=m3.data,
    start_date='',
    end_date='',
    before_start_days=0
)

m16 = M.derived_feature_extractor.v3(
    input_data=m15.data,
    features=m3.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False
)

m7 = M.join.v3(
    data1=m2.data,
    data2=m16.data,
    on='date,instrument',
    how='inner',
    sort=False
)

m13 = M.dropnan.v1(
    input_data=m7.data
)

m5 = M.stock_ranker_train.v6(
    training_ds=m13.data,
    features=m3.data,
    learning_algorithm='排序',
    number_of_leaves=30,
    minimum_docs_per_leaf=1000,
    number_of_trees=20,
    learning_rate=0.1,
    max_bins=1023,
    feature_fraction=1,
    data_row_fraction=1,
    plot_charts=True,
    ndcg_discount_base=1,
    m_lazy_run=False
)

m9 = M.instruments.v2(
    start_date=T.live_run_param('trading_date', '2015-01-01'),
    end_date=T.live_run_param('trading_date', '2017-01-01'),
    market='CN_STOCK_A',
    instrument_list='',
    max_count=0
)

m17 = M.general_feature_extractor.v7(
    instruments=m9.data,
    features=m3.data,
    start_date='',
    end_date='',
    before_start_days=60
)

m18 = M.derived_feature_extractor.v3(
    input_data=m17.data,
    features=m3.data,
    date_col='date',
    instrument_col='instrument',
    drop_na=False,
    remove_extra_columns=False
)

m14 = M.dropnan.v1(
    input_data=m18.data
)

m8 = M.stock_ranker_predict.v5(
    model=m5.model,
    data=m14.data,
    m_lazy_run=False
)

m4 = M.trade.v4(
    instruments=m9.data,
    options_data=m8.predictions,
    start_date='',
    end_date='',
    initialize=m4_initialize_bigquant_run,
    handle_data=m4_handle_data_bigquant_run,
    prepare=m4_prepare_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='后复权',
    product_type='股票',
    plot_charts=True,
    backtest_only=False,
    benchmark=''
)

[2021-12-01 11:32:37.549428] INFO: moduleinvoker: instruments.v2 开始运行..

[2021-12-01 11:32:37.573254] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.575437] INFO: moduleinvoker: instruments.v2 运行完成[0.026042s].

[2021-12-01 11:32:37.588778] INFO: moduleinvoker: advanced_auto_labeler.v2 开始运行..

[2021-12-01 11:32:37.608236] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.612691] INFO: moduleinvoker: advanced_auto_labeler.v2 运行完成[0.023909s].

[2021-12-01 11:32:37.622521] INFO: moduleinvoker: input_features.v1 开始运行..

[2021-12-01 11:32:37.634714] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.637347] INFO: moduleinvoker: input_features.v1 运行完成[0.014833s].

[2021-12-01 11:32:37.694836] INFO: moduleinvoker: general_feature_extractor.v7 开始运行..

[2021-12-01 11:32:37.710661] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.713813] INFO: moduleinvoker: general_feature_extractor.v7 运行完成[0.019001s].

[2021-12-01 11:32:37.735999] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2021-12-01 11:32:37.750140] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.753324] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[0.01733s].

[2021-12-01 11:32:37.772553] INFO: moduleinvoker: join.v3 开始运行..

[2021-12-01 11:32:37.787685] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.789987] INFO: moduleinvoker: join.v3 运行完成[0.017441s].

[2021-12-01 11:32:37.807194] INFO: moduleinvoker: dropnan.v1 开始运行..

[2021-12-01 11:32:37.824634] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:32:37.827139] INFO: moduleinvoker: dropnan.v1 运行完成[0.019993s].

[2021-12-01 11:32:37.849123] INFO: moduleinvoker: stock_ranker_train.v6 开始运行..

[2021-12-01 11:36:07.401267] INFO: StockRanker: 特征预处理 ..

[2021-12-01 11:36:11.619403] INFO: StockRanker: prepare data: training ..

[2021-12-01 11:38:10.559606] INFO: StockRanker: sort ..

[2021-12-01 11:41:22.913500] INFO: StockRanker训练: 54be91d2 准备训练: 2606084 行数

[2021-12-01 11:41:22.915071] INFO: StockRanker训练: AI模型训练，将在2606084*13=3387.91万数据上对模型训练进行20轮迭代训练。预计将需要11～21分钟。请耐心等待。

[2021-12-01 11:41:23.154240] INFO: StockRanker训练: 正在训练 ..

[2021-12-01 11:41:23.257509] INFO: StockRanker训练: 任务状态: Pending

[2021-12-01 11:41:33.336908] INFO: StockRanker训练: 任务状态: Running

[2021-12-01 11:41:53.528339] INFO: StockRanker训练: 00:00:19.5015738, finished iteration 1

[2021-12-01 11:42:13.696727] INFO: StockRanker训练: 00:00:33.9860826, finished iteration 2

[2021-12-01 11:42:23.802262] INFO: StockRanker训练: 00:00:48.6951035, finished iteration 3

[2021-12-01 11:42:43.995798] INFO: StockRanker训练: 00:01:05.5412082, finished iteration 4

[2021-12-01 11:42:54.072432] INFO: StockRanker训练: 00:01:21.3028336, finished iteration 5

[2021-12-01 11:43:14.214533] INFO: StockRanker训练: 00:01:41.9431485, finished iteration 6

[2021-12-01 11:43:44.368785] INFO: StockRanker训练: 00:02:05.5480695, finished iteration 7

[2021-12-01 11:44:04.570183] INFO: StockRanker训练: 00:02:27.8315558, finished iteration 8

[2021-12-01 11:44:24.748812] INFO: StockRanker训练: 00:02:51.1375583, finished iteration 9

[2021-12-01 11:44:44.960748] INFO: StockRanker训练: 00:03:14.3571939, finished iteration 10

[2021-12-01 11:45:15.194827] INFO: StockRanker训练: 00:03:37.1796134, finished iteration 11

[2021-12-01 11:45:35.366549] INFO: StockRanker训练: 00:04:01.0933458, finished iteration 12

[2021-12-01 11:45:55.555391] INFO: StockRanker训练: 00:04:24.2110812, finished iteration 13

[2021-12-01 11:46:25.800278] INFO: StockRanker训练: 00:04:47.4169547, finished iteration 14

[2021-12-01 11:46:45.929262] INFO: StockRanker训练: 00:05:10.7194360, finished iteration 15

[2021-12-01 11:47:06.072459] INFO: StockRanker训练: 00:05:34.6559775, finished iteration 16

[2021-12-01 11:47:36.316530] INFO: StockRanker训练: 00:05:58.5463097, finished iteration 17

[2021-12-01 11:47:56.595608] INFO: StockRanker训练: 00:06:22.9646795, finished iteration 18

[2021-12-01 11:48:16.817749] INFO: StockRanker训练: 00:06:41.5414397, finished iteration 19

[2021-12-01 11:48:36.912406] INFO: StockRanker训练: 00:07:01.4563048, finished iteration 20

[2021-12-01 11:48:36.914996] INFO: StockRanker训练: 任务状态: Succeeded

[2021-12-01 11:48:37.184895] INFO: moduleinvoker: stock_ranker_train.v6 运行完成[959.335768s].

[2021-12-01 11:48:37.194903] INFO: moduleinvoker: instruments.v2 开始运行..

[2021-12-01 11:48:37.216409] INFO: moduleinvoker: 命中缓存

[2021-12-01 11:48:37.218534] INFO: moduleinvoker: instruments.v2 运行完成[0.023634s].

[2021-12-01 11:48:37.259106] INFO: moduleinvoker: general_feature_extractor.v7 开始运行..

[2021-12-01 11:48:39.958660] INFO: 基础特征抽取: 年份 2014, 特征行数=99861

[2021-12-01 11:48:44.587416] INFO: 基础特征抽取: 年份 2015, 特征行数=569698

[2021-12-01 11:48:49.801307] INFO: 基础特征抽取: 年份 2016, 特征行数=641546

[2021-12-01 11:48:52.137946] INFO: 基础特征抽取: 年份 2017, 特征行数=0

[2021-12-01 11:48:52.306434] INFO: 基础特征抽取: 总行数: 1311105

[2021-12-01 11:48:52.312226] INFO: moduleinvoker: general_feature_extractor.v7 运行完成[15.053158s].

[2021-12-01 11:48:52.323644] INFO: moduleinvoker: derived_feature_extractor.v3 开始运行..

[2021-12-01 11:48:56.029760] INFO: derived_feature_extractor: 提取完成 avg_amount_0/avg_amount_5, 0.007s

[2021-12-01 11:48:56.038942] INFO: derived_feature_extractor: 提取完成 avg_amount_5/avg_amount_20, 0.006s

[2021-12-01 11:48:56.046486] INFO: derived_feature_extractor: 提取完成 rank_avg_amount_0/rank_avg_amount_5, 0.005s

[2021-12-01 11:48:56.055091] INFO: derived_feature_extractor: 提取完成 rank_avg_amount_5/rank_avg_amount_10, 0.004s

[2021-12-01 11:48:56.062511] INFO: derived_feature_extractor: 提取完成 rank_return_0/rank_return_5, 0.005s

[2021-12-01 11:48:56.069384] INFO: derived_feature_extractor: 提取完成 rank_return_5/rank_return_10, 0.004s

[2021-12-01 11:48:56.564574] INFO: derived_feature_extractor: /y_2014, 99861

[2021-12-01 11:48:58.054574] INFO: derived_feature_extractor: /y_2015, 569698

[2021-12-01 11:48:59.760753] INFO: derived_feature_extractor: /y_2016, 641546

[2021-12-01 11:49:00.437981] INFO: moduleinvoker: derived_feature_extractor.v3 运行完成[8.114321s].

[2021-12-01 11:49:00.452781] INFO: moduleinvoker: dropnan.v1 开始运行..

[2021-12-01 11:49:00.771365] INFO: dropnan: /y_2014, 99326/99861

[2021-12-01 11:49:01.836265] INFO: dropnan: /y_2015, 565146/569698

[2021-12-01 11:49:02.934333] INFO: dropnan: /y_2016, 636912/641546

[2021-12-01 11:49:03.076066] INFO: dropnan: 行数: 1301384/1311105

[2021-12-01 11:49:03.089558] INFO: moduleinvoker: dropnan.v1 运行完成[2.636766s].

[2021-12-01 11:49:03.114409] INFO: moduleinvoker: stock_ranker_predict.v5 开始运行..

[2021-12-01 11:49:03.864133] INFO: StockRanker预测: /y_2014 ..

[2021-12-01 11:49:04.770828] INFO: StockRanker预测: /y_2015 ..

[2021-12-01 11:49:16.470259] INFO: StockRanker预测: /y_2016 ..

[2021-12-01 11:49:28.853111] INFO: moduleinvoker: stock_ranker_predict.v5 运行完成[25.738698s].

[2021-12-01 11:49:30.943311] INFO: moduleinvoker: backtest.v8 开始运行..

[2021-12-01 11:49:30.949422] INFO: backtest: biglearning backtest:V8.6.0

[2021-12-01 11:49:30.951502] INFO: backtest: product_type:stock by specified

[2021-12-01 11:49:31.114890] INFO: moduleinvoker: cached.v2 开始运行..

[2021-12-01 11:49:55.748338] INFO: backtest: 读取股票行情完成:2212017

[2021-12-01 11:50:00.586754] INFO: moduleinvoker: cached.v2 运行完成[29.471877s].

[2021-12-01 11:50:03.094995] INFO: algo: TradingAlgorithm V1.8.5

[2021-12-01 11:50:04.188416] INFO: algo: trading transform...

[2021-12-01 11:50:22.181178] INFO: Performance: Simulated 488 trading days out of 488.

[2021-12-01 11:50:22.182910] INFO: Performance: first open: 2015-01-05 09:30:00+00:00

[2021-12-01 11:50:22.184988] INFO: Performance: last close: 2016-12-30 15:00:00+00:00

[2021-12-01 11:50:30.757929] INFO: moduleinvoker: backtest.v8 运行完成[59.81462s].

[2021-12-01 11:50:30.760553] INFO: moduleinvoker: trade.v4 运行完成[61.893966s].

设置评估测试数据集，查看训练曲线
[视频教程]StockRanker训练曲线

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-0aa2f3e952d744fdbd3c7731dc2adf42"}/bigcharts-data-end

收益率318.76%
年化收益率109.5%
基准收益率-6.33%
阿尔法1.24
贝塔0.93
夏普比率1.93
胜率0.62
盈亏比0.94
收益波动率41.31%
信息比率0.18
最大回撤47.62%

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-32ae6f9dba0644bea1bdcf97e10588bd"}/bigcharts-data-end

In [16]:

m5

Out[16]:

{'version': 'v3', 'test_ds': None, 'model_id': '54be91d2525711eca6571e22cc2cc31c', 'model': DataSource(db8b34b64d6645459dd1b5c358377302T), 'plot_charts': True, 'test_ndcg_ds': None, 'feature_gains': DataSource(c760182c89c842bbb569c2c4bbc63f96T), 'plot_feature_gains': <bound method plot_feature_gains of {...}>, 'plot_test_data_ndcg': <bound method plot_test_data_ndcg of {...}>, 'plot_model': <bound method plot_model of {...}>, 'plot_stock_ranker_tabs': <bound method plot_stock_ranker_tabs of {...}>}

In [27]:

m5.model_id

Out[27]:

'54be91d2525711eca6571e22cc2cc31c'

In [19]:

m5.feature_gains.read()

Out[19]:

	feature	gain
0	rank_avg_amount_5/rank_avg_amount_10	650.052263
1	rank_return_5	417.164155
2	rank_return_0	283.331818
3	return_20	165.323698
4	rank_avg_amount_0/rank_avg_amount_5	148.071940
5	rank_return_10	99.256969
6	rank_return_0/rank_return_5	65.245423
7	rank_return_5/rank_return_10	65.130339
8	pe_ttm_0	39.884205
9	return_10	27.107671
10	avg_amount_5/avg_amount_20	19.906750
11	avg_amount_0/avg_amount_5	16.892008
12	return_5	10.942141

In [8]:

m5.plot_stock_ranker_tabs()

设置评估测试数据集，查看训练曲线
[视频教程]StockRanker训练曲线

bigcharts-data-start/{"__type":"tabs","__id":"bigchart-a88bdb44505e49a7a41092d0f7d58ba4"}/bigcharts-data-end