
Stock Index Prediction and Industry Rotation Strategy Based on a Decomposition-Reconstruction-Integration Framework
Li ZHENG, Mingchen LI, Yunjie WEI, Shouyang WANG
China Journal of Econometrics ›› 2024, Vol. 4 ›› Issue (3) : 673-698.
Stock Index Prediction and Industry Rotation Strategy Based on a Decomposition-Reconstruction-Integration Framework
Stock index data is influenced by multiple factors, exhibiting nonlinear, non-stationary, high complexity, and high volatility characteristics. Therefore, it is difficult for a single model to fully capture its data features. This paper proposes a hybrid forecasting model for stock index returns based on a decomposition-reconstruction-integration framework. Utilizing variational mode decomposition (VMD), the original high-complexity stock index time series is decomposed. The composite multiscale entropy (CMSE) is employed as a reconstruction indicator to reorganize the stock index data components into long-term trend terms, medium-term impact terms, and short-term disturbance terms. ARIMA, BPNN, and LSTM models are adopted for forecasting based on their respective data characteristics. Finally, the predictions of various frequency components are integrated to obtain the ultimate forecasting results. The proposed method is applied to forecasting eight significant industry stock indices and is compared with models utilizing fine to coarse (FTC), sample entropy (SE), fuzzy entropy (FE), and multiscale permutation entropy (MSPE) as reconstruction methods. Furthermore, two industry rotation strategies, equal-weight investment and dynamic-weight investment, are proposed to validate the performance of the proposed model in practical trading from both conservative and aggressive perspectives. The empirical results demonstrate that CMSE outperforms other reconstruction indicators in stock index forecasting. Compared to benchmark models, the hybrid model presented in this paper achieves lower forecasting errors and higher directional accuracy, and the proposed industry rotation strategies exhibit excellent performance in terms of risk and return.
decomposition-reconstruction-integration / stock index prediction / composite multiscale entropy / industry rotation {{custom_keyword}} /
表1 参数设置 |
模型/算法 | 设置 |
ARIMA | R语言, “forecast'' 包中的auto.arima()函数 |
BPNN | hidden layer sizes=(16, 4), activation=relu, solver=adam, alpha=0.01, batch size=1, epochs=100 |
LSTM | LSTM units=16, optimizer=adam, loss=mean squared error, learning rate=0.001, batch size=1, epochs=100 |
VMD | alpha = 1000, |
表2 重构结果 |
EMD | FTC | 样本熵 | 模糊熵 | 多尺度排列熵 | 复合多尺度熵 |
长期趋势项 | Residual | Residual, IMF1 | Residual, IMF1 | Residual, IMF1 | Residual, IMF1 |
中期影响项 | IMF1 | IMF4 | IMF5 | IMF5 | IMF6 |
短期扰动项 | IMF4 | IMF7 | IMF7 | IMF8 | IMF8 |
VMD | FTC | 样本熵 | 模糊熵 | 多尺度排列熵 | 复合多尺度熵 |
长期趋势项 | Residual | Residual, IMF1 | Residual, IMF1 | Residual, IMF1 | Residual, IMF1 |
中期影响项 | None | IMF3 | IMF3 | IMF5 | IMF4 |
短期扰动项 | IMF1 | IMF5 | IMF6 | IMF6 | IMF6 |
表3 预测误差比较 |
基础化工 | RMSE | MAPE | DA | 电力设备 | RMSE | MAPE | DA | |||
ARIMA | 75.6395 | 0.0138 | 0.4953 | 0.3558 | ARIMA | 223.2676 | 0.0183 | 0.4933 | 0.3356 | |
BPNN | 60.8449 | 0.0115 | 0.5180 | 0.5580 | BPNN | 187.5506 | 0.0157 | 0.4881 | 0.5616 | |
LSTM | 61.2032 | 0.0114 | 0.5190 | 0.6361 | LSTM | 194.1033 | 0.0160 | 0.4912 | 0.5316 | |
EMD-FTC | 49.2360 | 0.0093 | 0.6474 | 0.6640 | EMD-FTC | 133.8422 | 0.0118 | 0.6484 | 0.6538 | |
EMD-SE | 48.1886 | 0.0100 | 0.6556 | 0.6692 | EMD-SE | 137.5171 | 0.0120 | 0.6556 | 0.6584 | |
EMD-FE | 46.8049 | 0.0102 | 0.7072 | 0.7277 | EMD-FE | 134.7100 | 0.0115 | 0.6608 | 0.6639 | |
EMD-MSPE | 45.8271 | 0.0087 | 0.7164 | 0.7363 | EMD-MSPE | 129.1808 | 0.0109 | 0.6845 | 0.6927 | |
EMD-CMSE | 43.8558 | 0.0091 | 0.7000 | 0.7166 | EMD-CMSE | 130.9736 | 0.0112 | 0.6783 | 0.6835 | |
VMD-FTC | 48.5146 | 0.0104 | 0.6752 | 0.6853 | VMD-FTC | 132.4198 | 0.0129 | 0.6711 | 0.6734 | |
VMD-SE | 44.7810 | 0.0091 | 0.6786 | 0.6929 | VMD-SE | 132.0359 | 0.0123 | 0.6622 | 0.6686 | |
VMD-FE | 42.9091 | 0.0086 | 0.7116 | 0.7222 | VMD-FE | 127.8720 | 0.0119 | 0.6714 | 0.6800 | |
VMD-MSPE | 41.0506 | 0.0082 | 0.7322 | 0.7435 | VMD-MSPE | 125.7633 | 0.0116 | 0.6694 | 0.6741 | |
VMD-CMSE | 40.8689 | 0.0077 | 0.7229 | 0.7385 | VMD-CMSE | 120.9940 | 0.0100 | 0.6910 | 0.7047 | |
机械设备 | RMSE | MAPE | DA | 汽车 | RMSE | MAPE | DA | |||
ARIMA | 24.9460 | 0.0128 | 0.4902 | 0.3261 | ARIMA | 116.6076 | 0.0158 | 0.5231 | 0.3441 | |
BPNN | 20.3051 | 0.0105 | 0.5097 | 0.4994 | BPNN | 93.3247 | 0.0129 | 0.5272 | 0.5719 | |
LSTM | 20.4840 | 0.0105 | 0.4902 | 0.6962 | LSTM | 93.1862 | 0.0129 | 0.5231 | 0.6358 | |
EMD-FTC | 16.7632 | 0.0089 | 0.6762 | 0.6945 | EMD-FTC | 68.6535 | 0.0097 | 0.6659 | 0.6713 | |
EMD-SE | 17.7022 | 0.0094 | 0.6804 | 0.7007 | EMD-SE | 72.3501 | 0.0103 | 0.6474 | 0.6510 | |
EMD-FE | 16.9772 | 0.0093 | 0.6865 | 0.7013 | EMD-FE | 73.6916 | 0.0107 | 0.6432 | 0.6484 | |
EMD-MSPE | 16.0392 | 0.0083 | 0.6567 | 0.6794 | EMD-MSPE | 70.4071 | 0.0102 | 0.6525 | 0.6550 | |
EMD-CMSE | 15.0742 | 0.0078 | 0.6793 | 0.6995 | EMD-CMSE | 66.6971 | 0.0096 | 0.6639 | 0.6733 | |
VMD-FTC | 19.1389 | 0.0110 | 0.6752 | 0.6878 | VMD-FTC | 82.6860 | 0.01248 | 0.6731 | 0.6807 | |
VMD-SE | 17.1356 | 0.0094 | 0.6313 | 0.6404 | VMD-SE | 72.1996 | 0.0109 | 0.6416 | 0.6506 | |
VMD-FE | 15.8812 | 0.0087 | 0.6477 | 0.6660 | VMD-FE | 68.4636 | 0.0102 | 0.6601 | 0.6646 | |
VMD-MSPE | 15.9036 | 0.0086 | 0.6481 | 0.6481 | VMD-MSPE | 68.5173 | 0.0102 | 0.6549 | 0.6633 | |
VMD-CMSE | 13.6513 | 0.0071 | 0.6333 | 0.6563 | VMD-CMSE | 64.7736 | 0.0091 | 0.6889 | 0.6986 | |
食品饮料 | RMSE | MAPE | DA | 医药生物 | RMSE | MAPE | DA | |||
ARIMA | 520.8553 | 0.0163 | 0.4974 | 0.3761 | ARIMA | 206.9296 | 0.0147 | 0.4850 | 0.3242 | |
BPNN | 425.4355 | 0.0137 | 0.5221 | 0.4940 | BPNN | 170.8959 | 0.0125 | 0.4881 | 0.4529 | |
LSTM | 419.0281 | 0.0137 | 0.5066 | 0.5188 | LSTM | 181.4243 | 0.0130 | 0.5087 | 0.4633 | |
EMD-FTC | 352.2843 | 0.0118 | 0.6062 | 0.6133 | EMD-FTC | 151.9797 | 0.0121 | 0.6814 | 0.6894 | |
EMD-SE | 363.9338 | 0.0125 | 0.6021 | 0.6045 | EMD-SE | 154.8763 | 0.0123 | 0.6659 | 0.6760 | |
EMD-FE | 360.5713 | 0.0124 | 0.6237 | 0.6309 | EMD-FE | 157.1087 | 0.0125 | 0.6484 | 0.6552 | |
EMD-MSPE | 354.0642 | 0.0122 | 0.6247 | 0.6262 | EMD-MSPE | 151.6850 | 0.0120 | 0.6515 | 0.6536 | |
EMD-CMSE | 350.3146 | 0.0120 | 0.6329 | 0.6322 | EMD-CMSE | 147.9094 | 0.0118 | 0.6855 | 0.6897 | |
VMD-FTC | 375.0199 | 0.01341 | 0.6597 | 0.6639 | VMD-FTC | 150.0258 | 0.0122 | 0.6742 | 0.6755 | |
VMD-SE | 359.6157 | 0.0128 | 0.6591 | 0.6577 | VMD-SE | 148.1139 | 0.0121 | 0.6722 | 0.6774 | |
VMD-FE | 363.7681 | 0.0130 | 0.6323 | 0.6330 | VMD-FE | 145.4940 | 0.0119 | 0.6711 | 0.6735 | |
VMD-MSPE | 354.5632 | 0.0124 | 0.6488 | 0.6552 | VMD-MSPE | 138.0762 | 0.0113 | 0.6721 | 0.6794 | |
VMD-CMSE | 332.5511 | 0.0118 | 0.6601 | 0.6597 | VMD-CMSE | 132.9015 | 0.0104 | 0.6907 | 0.6963 | |
电子 | RMSE | MAPE | DA | 计算机 | RMSE | MAPE | DA | |||
ARIMA | 96.4673 | 0.0171 | 0.5200 | 0.3575 | ARIMA | 108.8869 | 0.0166 | 0.4819 | 0.3451 | |
BPNN | 80.7064 | 0.0143 | 0.5221 | 0.6558 | BPNN | 92.9928 | 0.0141 | 0.5046 | 0.4853 | |
LSTM | 78.4350 | 0.0139 | 0.5118 | 0.5784 | LSTM | 90.5214 | 0.0138 | 0.4974 | 0.5387 | |
EMD-FTC | 73.6436 | 0.0142 | 0.6443 | 0.6511 | EMD-FTC | 68.8637 | 0.0114 | 0.6453 | 0.6446 | |
EMD-SE | 75.9693 | 0.0147 | 0.6319 | 0.6390 | EMD-SE | 74.3434 | 0.0124 | 0.6268 | 0.6306 | |
EMD-FE | 74.2275 | 0.0142 | 0.6329 | 0.6418 | EMD-FE | 72.9007 | 0.0123 | 0.6422 | 0.6404 | |
EMD-MSPE | 72.5315 | 0.0139 | 0.6288 | 0.6341 | EMD-MSPE | 74.3527 | 0.0122 | 0.6360 | 0.6342 | |
EMD-CMSE | 70.2439 | 0.0135 | 0.6360 | 0.6445 | EMD-CMSE | 68.1861 | 0.0112 | 0.6494 | 0.6551 | |
VMD-FTC | 76.1730 | 0.0148 | 0.6391 | 0.6492 | VMD-FTC | 74.5698 | 0.0127 | 0.6453 | 0.6504 | |
VMD-SE | 75.3931 | 0.0145 | 0.6567 | 0.6659 | VMD-SE | 73.3949 | 0.0120 | 0.6148 | 0.6160 | |
VMD-FE | 72.4956 | 0.0140 | 0.6505 | 0.6579 | VMD-FE | 72.9063 | 0.0119 | 0.6251 | 0.6216 | |
VMD-MSPE | 68.7628 | 0.0133 | 0.6649 | 0.6713 | VMD-MSPE | 72.8293 | 0.0119 | 0.6302 | 0.6279 | |
VMD-CMSE | 66.2225 | 0.0127 | 0.6927 | 0.7025 | VMD-CMSE | 64.4893 | 0.0100 | 0.6498 | 0.6509 |
表4 DM检验结果(以VMD-CMSE为基准模型) |
基础化工 | 电力设备 | 机械设备 | 汽车 | 食品饮料 | 医药生物 | 电子 | 计算机 | |
ARIMA | 9.59*** | 11.50*** | 11.60*** | 11.22*** | 7.98*** | 8.55*** | 8.00*** | 9.51*** |
BPNN | 8.52*** | 9.55*** | 8.34*** | 8.41*** | 5.84*** | 6.35*** | 4.92*** | 7.21*** |
LSTM | 8.06*** | 9.97*** | 8.24*** | 8.50*** | 5.48*** | 7.40*** | 4.26*** | 7.10*** |
EMD-FTC | 4.65*** | 2.75**** | 5.49*** | 1.48 | 1.60 | 4.48*** | 7.73*** | 1.84* |
EMD-SE | 5.00*** | 3.48*** | 6.46*** | 2.89*** | 2.56** | 5.02*** | 8.23*** | 3.92*** |
EMD-FE | 3.57*** | 2.89*** | 6.00*** | 3.77*** | 2.20** | 5.42*** | 3.66*** | 3.41*** |
EMD-MSPE | 3.25*** | 1.77* | 4.01*** | 2.45** | 1.70* | 4.11*** | 2.95*** | 3.99*** |
EMD-CMSE | 2.08** | 2.48** | 2.50** | 0.83 | 1.42 | 3.32*** | 1.97** | 1.59 |
VMD-FTC | 5.05*** | 2.55** | 10.18*** | 7.42*** | 3.74*** | 3.80*** | 4.57*** | 4.14*** |
VMD-SE | 3.24*** | 3.25*** | 7.66*** | 4.03*** | 4.23*** | 3.41*** | 3.95*** | 4.60*** |
VMD-FE | 1.66* | 2.06** | 5.47*** | 2.04** | 4.35*** | 2.84*** | 2.97*** | 4.45*** |
VMD-MSPE | 0.15 | 1.41 | 5.27*** | 1.90* | 3.11*** | 1.21 | 1.22 | 3.97*** |
注: * 代表结果在0.1的显著性水平下显著; ** 代表结果在0.05的显著性水平下显著; *** 代表结果在0.01的显著性水平下显著. |
表5 复合模型相较于LSTM模型的平均优化百分比 |
RMSE | MAPE | DA | ||
EMD-FTC | 24.44 | 18.32 | 28.87 | 14.88 |
EMD-SE | 20.54 | 12.80 | 27.63 | 13.73 |
EMD-FE | 21.56 | 13.40 | 32.60 | 15.47 |
EMD-MSPE | 24.60 | 19.21 | 29.78 | 15.52 |
EMD-CMSE | 27.51 | 22.45 | 31.63 | 17.32 |
VMD-FTC | 18.82 | 5.80 | 31.26 | 16.71 |
VMD-SE | 23.40 | 13.01 | 28.89 | 14.60 |
VMD-FE | 25.14 | 16.71 | 30.20 | 15.67 |
VMD-MSPE | 28.58 | 20.22 | 31.12 | 16.63 |
VMD-CMSE | 36.11 | 33.60 | 34.15 | 19.78 |
表6 使用同一重构方法时VMD分解相较于EMD分解预测的平均优化百分比 |
RMSE | MAPE | DA | ||
FTC | 1.82 | 1.57 | ||
SE | 2.37 | 0.19 | 0.98 | 0.76 |
FE | 2.95 | 2.92 | 0.17 | |
MSPE | 3.19 | 0.85 | 1.02 | 0.95 |
CMSE | 6.75 | 9.11 | 1.88 | 2.05 |
注: 第一行代表对于八个行业指数的预测, VMD-FTC模型相较于EMD-FTC模型的预测结果在 |
表7 不同模型及策略的交易表现 |
策略一 | 累计收益(%) | 年化收益(%) | 下行风险(%) | 最大回撤(%) | 夏普比率 | 索提诺比率 |
EMD-FTC | 383.16 | 50.56 | 0.1451 | 24.62 | 1.74 | 2.73 |
EMD-SE | 266.12 | 40.09 | 0.1476 | 28.21 | 1.42 | 2.20 |
EMD-FE | 343.97 | 47.29 | 0.1462 | 24.10 | 1.64 | 2.57 |
EMD-MSPE | 450.58 | 55.76 | 0.1448 | 22.23 | 1.89 | 2.97 |
EMD-CMSE | 507.21 | 59.77 | 0.1452 | 20.40 | 1.99 | 3.14 |
VMD-FTC | 436.99 | 54.75 | 0.1511 | 24.45 | 1.82 | 2.82 |
VMD-SE | 367.59 | 49.28 | 0.1518 | 24.88 | 1.66 | 2.57 |
VMD-FE | 439.41 | 54.93 | 0.1512 | 24.00 | 1.83 | 2.83 |
VMD-MSPE | 373.03 | 49.73 | 0.1509 | 24.22 | 1.68 | 2.60 |
VMD-CMSE | 633.93 | 67.83 | 0.1465 | 24.88 | 2.19 | 3.45 |
BH* | 81.37 | 16.73 | 0.2153 | 49.45 | 0.49 | 0.79 |
策略二 | 累计收益(%) | 年化收益(%) | 下行风险 | 最大回撤(%) | 夏普比率 | 索提诺比率 |
EMD-FTC | 360.74 | 48.72 | 0.1589 | 28.62 | 1.58 | 2.45 |
EMD-SE | 289.06 | 42.33 | 0.1602 | 28.49 | 1.40 | 2.15 |
EMD-FE | 386.02 | 50.79 | 0.1555 | 22.07 | 1.65 | 2.58 |
EMD-MSPE | 662.13 | 69.49 | 0.1544 | 19.87 | 2.12 | 3.36 |
EMD-CMSE | 777.17 | 75.79 | 0.1547 | 20.25 | 2.25 | 3.59 |
VMD-FTC | 722.38 | 72.87 | 0.1617 | 26.43 | 2.17 | 3.34 |
VMD-SE | 549.98 | 62.62 | 0.1607 | 26.32 | 1.93 | 2.98 |
VMD-FE | 567.94 | 63.78 | 0.1581 | 27.01 | 1.96 | 3.07 |
VMD-MSPE | 499.14 | 59.22 | 0.1593 | 27.70 | 1.86 | 2.87 |
VMD-CMSE | 865.97 | 80.25 | 0.1547 | 26.80 | 2.35 | 3.75 |
BH* | 81.37 | 16.73 | 0.2153 | 49.45 | 0.49 | 0.79 |
陈坚, 张轶凡, 中国股票市场的已实现偏度与收益率预测[J]. 金融研究, 2018,(9): 107- 125.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
陈凯杰, 唐振鹏, 吴俊传, 杜晓旭, 蔡毅, 基于分解-集成和混频数据采样的中国股票市场预测研究[J]. 系统工程理论与实践, 2022, 42(11): 3105- 3120.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
邓可斌, 关子桓, 陈彬, 宏观经济政策与股市系统性风险——宏微观混合
{{custom_citation.content}}
{{custom_citation.annotation}}
|
高海翔, 胡瑜, 余乐安, 基于分解集成的LSTM神经网络模型的油价预测[J]. 计算机应用与软件, 2021, 38, 78- 83.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
姜富伟, 郭鹏, 郭豫媚, 美联储货币政策对我国资产价格的影响[J]. 金融研究, 2019,(5): 37- 55.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
李岸, 粟亚亚, 乔海曙, 中国股票市场国际联动性研究——基于网络分析方法[J]. 数量经济技术经济研究, 2016, 33(8): 113- 127.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
李瑞, 范玉刚, 基于CEEMDAN多尺度排列熵和SO-RELM的高压隔膜泵单向阀故障诊断[J]. 振动与冲击, 2023, 42(5): 127- 135.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
梁小珍, 郭战坤, 张倩文, 杨明歌, 汪寿阳, 基于奇异谱分析的航空客运需求分析与分解集成预测模型[J]. 系统工程理论与实践, 2020, 40(7): 1844- 1855.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
万建强, 文洲, ARIMA模型与ARCH模型在香港股指预测方面的应用比较[J]. 数理统计与管理, 2001, 20(6): 1- 4.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
闻岳春, 王婧婷, 基于FA-BPNN的金融控股公司风险预警系统研究——以美国和台湾地区为例[J]. 国际金融研究, 2010,(11): 37- 46.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
许年行, 于上尧, 伊志宏, 机构投资者羊群行为与股价崩盘风险[J]. 管理世界, 2013, 7(3): 31- 43.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
杨丽, 吴雨茜, 王俊丽, 刘义理, 循环神经网络研究综述[J]. 计算机应用, 2018, 38(A2): 1- 6.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
朱民, 徐钟祥, 中美贸易摩擦对中国股票市场的影响[J]. 国际金融研究, 2021,(4): 3- 12.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
Aziz W, Arif M, (2005). Multiscale Permutation Entropy of Physiological Time Series[C]// 2005 Pakistan Section Multitopic Conference, IEEE.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |