克隆策略
In [6]:
# 基础参数配置
class conf:
    start_date = '2010-01-01'
    end_date='2017-07-01'
    # split_date 之前的数据用于训练,之后的数据用作效果评估
    split_date = '2015-01-01'
    # D.instruments: https://bigquant.com/docs/data_instruments.html
    instruments = D.instruments(start_date, split_date)

    # 机器学习目标标注函数
    # 如下标注函数等价于 min(max((持有期间的收益 * 100), -20), 20) + 20 (后面的M.fast_auto_labeler会做取整操作)
    # 说明:max/min这里将标注分数限定在区间[-20, 20],+20将分数变为非负数 (StockRanker要求标注分数非负整数)
    label_expr = ['return * 100', 'where(label > {0}, {0}, where(label < -{0}, -{0}, label)) + {0}'.format(20)]
    # 持有天数,用于计算label_expr中的return值(收益)
    hold_days = 3

    # 特征 https://bigquant.com/docs/data_features.html,你可以通过表达式构造任何特征
    features = [
        'ta_sma_5_0',  # 5日移动平均
#         'ta_sma_10_0',  # 10日移动平均
        'ta_sma_20_0',  # 20日移动平均
#         'ta_mom_10',  # 10日动量
#         'ta_mom_20',  # 20日动量
#         'ta_mom_30',  # 30日动量
#         'ta_mom_60',  # 60日动量
#         'ta_atr_14_0',  # 14日平均真实波幅
#         'ta_atr_28_0',  # 28日平均真实波幅
#         'ta_mfi_14_0',  # 14日资金流量指标
#         'ta_mfi_28_0',  # 28日资金流量指标
#         'ta_rsi_14_0',  # 14日相对强弱指标
#         'ta_rsi_28_0',  # 28日相对强弱指标
#         'ta_willr_14_0',  # 14日威廉指标
#         'ta_willr_28_0',  # 28日威廉指标
    ]

# 给数据做标注:给每一行数据(样本)打分,一般分数越高表示越好
m1 = M.fast_auto_labeler.v5(
    instruments=conf.instruments, start_date=conf.start_date, end_date=conf.split_date,
    label_expr=conf.label_expr, hold_days=conf.hold_days,
    benchmark='000300.SHA', sell_at='open', buy_at='open')
# 计算特征数据
m2 = M.general_feature_extractor.v5(
    instruments=conf.instruments, start_date=conf.start_date, end_date=conf.split_date,
    features=conf.features)
# 数据预处理:缺失数据处理,数据规范化,T.get_stock_ranker_default_transforms为StockRanker模型做数据预处理
m3 = M.transform.v2(
    data=m2.data, transforms=T.get_stock_ranker_default_transforms(),
    drop_null=True, astype='int32', except_columns=['date', 'instrument'],
    clip_lower=0, clip_upper=200000000)
# 合并标注和特征数据
m4 = M.join.v2(data1=m1.data, data2=m3.data, on=['date', 'instrument'], sort=True)
# StockRanker机器学习训练
m5 = M.stock_ranker_train.v2(training_ds=m4.data, features=conf.features)


## 量化回测 https://bigquant.com/docs/module_trade.html
# 回测引擎:准备数据,只执行一次
def prepare(context):
    # context.start_date / end_date,回测的时候,为trader传入参数;在实盘运行的时候,由系统替换为实盘日期
    n1 = M.general_feature_extractor.v5(
        instruments=D.instruments(),
        start_date=context.start_date, end_date=context.end_date,
        model_id=context.options['model_id'])
    n2 = M.transform.v2(
        data=n1.data, transforms=T.get_stock_ranker_default_transforms(),
        drop_null=True, astype='int32', except_columns=['date', 'instrument'],
        clip_lower=0, clip_upper=200000000)
    n3 = M.stock_ranker_predict.v2(model_id=context.options['model_id'], data=n2.data)
    context.instruments = n3.instruments
    context.options['predictions'] = n3.predictions

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

# 回测引擎:每日数据处理函数,每天执行一次
def handle_data(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StockRanker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StockRanker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)


# 调用交易引擎
m6 = M.trade.v1(
    instruments=None,
    start_date=conf.split_date,
    end_date=conf.end_date,
    prepare=prepare,
    initialize=initialize,
    handle_data=handle_data,
    order_price_field_buy='open',       # 表示 开盘 时买入
    order_price_field_sell='close',     # 表示 收盘 前卖出
    capital_base=1000000,               # 初始资金
    benchmark='000300.SHA',             # 比较基准,不影响回测结果
    # 通过 options 参数传递预测数据和参数给回测引擎
    options={'hold_days': conf.hold_days, 'model_id': m5.model_id}
)
[2017-07-17 13:51:31.449165] WARNING: bigquant: 此模块版本 M.fast_auto_labeler.v5 已不再维护,并可能在未来被删除:请更新到 fast_auto_labeler 最新版本
[2017-07-17 13:51:31.450897] INFO: bigquant: fast_auto_labeler.v5 start ..
[2017-07-17 13:51:31.452787] INFO: bigquant: hit cache
[2017-07-17 13:51:31.456687] INFO: bigquant: fast_auto_labeler.v5 end [0.005783s].
[2017-07-17 13:51:31.460987] WARNING: bigquant: 此模块版本 M.general_feature_extractor.v5 已不再维护,并可能在未来被删除:请更新到 general_feature_extractor 最新版本
[2017-07-17 13:51:31.462368] INFO: bigquant: general_feature_extractor.v5 start ..
[2017-07-17 13:51:31.880269] INFO: general_feature_extractor: year 2010, featurerows=431567
[2017-07-17 13:51:32.369460] INFO: general_feature_extractor: year 2011, featurerows=511455
[2017-07-17 13:51:32.915624] INFO: general_feature_extractor: year 2012, featurerows=565675
[2017-07-17 13:51:33.523443] INFO: general_feature_extractor: year 2013, featurerows=564168
[2017-07-17 13:51:34.105181] INFO: general_feature_extractor: year 2014, featurerows=569948
[2017-07-17 13:51:34.398961] INFO: general_feature_extractor: year 2015, featurerows=0
[2017-07-17 13:51:34.412396] INFO: general_feature_extractor: total feature rows: 2642813
[2017-07-17 13:51:34.414369] INFO: bigquant: general_feature_extractor.v5 end [2.951995s].
[2017-07-17 13:51:34.421326] INFO: bigquant: transform.v2 start ..
[2017-07-17 13:51:34.962940] INFO: transform: transformed /y_2010, 424955/431567
[2017-07-17 13:51:35.459356] INFO: transform: transformed /y_2011, 505992/511455
[2017-07-17 13:51:35.991378] INFO: transform: transformed /y_2012, 562552/565675
[2017-07-17 13:51:36.523342] INFO: transform: transformed /y_2013, 564140/564168
[2017-07-17 13:51:37.069680] INFO: transform: transformed /y_2014, 567742/569948
[2017-07-17 13:51:37.085079] INFO: transform: transformed rows: 2625381/2642813
[2017-07-17 13:51:37.105146] INFO: bigquant: transform.v2 end [2.683794s].
[2017-07-17 13:51:37.111128] INFO: bigquant: join.v2 start ..
[2017-07-17 13:51:39.500550] INFO: join: /y_2010, rows=424427/424955, timetaken=1.818395s
[2017-07-17 13:51:41.453078] INFO: join: /y_2011, rows=505466/505992, timetaken=1.928889s
[2017-07-17 13:51:43.518577] INFO: join: /y_2012, rows=561462/562552, timetaken=2.030708s
[2017-07-17 13:51:45.673334] INFO: join: /y_2013, rows=563109/564140, timetaken=2.117056s
[2017-07-17 13:51:47.836399] INFO: join: /y_2014, rows=556175/567742, timetaken=2.118907s
[2017-07-17 13:51:47.938843] INFO: join: total result rows: 2610639
[2017-07-17 13:51:47.941029] INFO: bigquant: join.v2 end [10.829891s].
[2017-07-17 13:51:47.947057] WARNING: bigquant: 此模块版本 M.stock_ranker_train.v2 已不再维护,并可能在未来被删除:请更新到 stock_ranker_train 最新版本
[2017-07-17 13:51:47.947919] INFO: bigquant: stock_ranker_train.v2 start ..
[2017-07-17 13:51:47.952584] WARNING: bigquant: 此模块版本 M.cached.v1 已不再维护,并可能在未来被删除:请更新到 cached 最新版本
[2017-07-17 13:51:49.285660] INFO: df2bin: prepare data: training ..
[2017-07-17 13:52:14.288318] INFO: stock_ranker_train: training: 2610639 rows
[2017-07-17 13:54:17.720388] INFO: bigquant: stock_ranker_train.v2 end [149.772403s].
[2017-07-17 13:54:17.738401] INFO: bigquant: backtest.v6 start ..
[2017-07-17 13:54:17.845909] WARNING: bigquant: 此模块版本 M.general_feature_extractor.v5 已不再维护,并可能在未来被删除:请更新到 general_feature_extractor 最新版本
[2017-07-17 13:54:17.847936] INFO: bigquant: general_feature_extractor.v5 start ..
[2017-07-17 13:54:19.901620] INFO: general_feature_extractor: year 2015, featurerows=569698
[2017-07-17 13:54:22.206945] INFO: general_feature_extractor: year 2016, featurerows=641546
[2017-07-17 13:54:23.630766] INFO: general_feature_extractor: year 2017, featurerows=349120
[2017-07-17 13:54:23.651318] INFO: general_feature_extractor: total feature rows: 1560364
[2017-07-17 13:54:23.657546] INFO: bigquant: general_feature_extractor.v5 end [5.809581s].
[2017-07-17 13:54:23.664674] INFO: bigquant: transform.v2 start ..
[2017-07-17 13:54:24.276946] INFO: transform: transformed /y_2015, 565564/569698
[2017-07-17 13:54:24.842906] INFO: transform: transformed /y_2016, 637338/641546
[2017-07-17 13:54:25.229529] INFO: transform: transformed /y_2017, 344310/349120
[2017-07-17 13:54:25.238903] INFO: transform: transformed rows: 1547212/1560364
[2017-07-17 13:54:25.249172] INFO: bigquant: transform.v2 end [1.584485s].
[2017-07-17 13:54:25.254744] INFO: bigquant: stock_ranker_predict.v2 start ..
[2017-07-17 13:54:25.259455] WARNING: bigquant: 此模块版本 M.cached.v1 已不再维护,并可能在未来被删除:请更新到 cached 最新版本
[2017-07-17 13:54:25.791406] INFO: df2bin: prepare data: prediction ..
[2017-07-17 13:54:40.266713] INFO: stock_ranker_predict: prediction: 1547212 rows
[2017-07-17 13:54:49.589839] INFO: bigquant: stock_ranker_predict.v2 end [24.335056s].
[2017-07-17 13:55:20.307821] INFO: Performance: Simulated 607 trading days out of 607.
[2017-07-17 13:55:20.308954] INFO: Performance: first open: 2015-01-05 14:30:00+00:00
[2017-07-17 13:55:20.309815] INFO: Performance: last close: 2017-06-30 19:00:00+00:00
[注意] 有 1 笔卖出是在多天内完成的。当日卖出股票超过了当日股票交易的2.5%会出现这种情况。
  • 收益率150.24%
  • 年化收益率46.35%
  • 基准收益率3.77%
  • 阿尔法0.45
  • 贝塔0.97
  • 夏普比率1.05
  • 收益波动率40.11%
  • 信息比率1.56
  • 最大回撤50.75%
[2017-07-17 13:55:22.434971] INFO: bigquant: backtest.v6 end [64.696557s].