學術產出-Theses

Article View/Open

Publication Export

Google ScholarTM

政大圖書館

Citation Infomation

題名 強化學習下動態調整風險偏好之投資組合配置:以台灣50指數為例
Portfolio Allocation with Dynamic Risk Aversion via Reinforcement Learning: Evidence from Taiwan 50 Index
作者 陳昱成
Chen, Yu-Cheng
貢獻者 林士貴
陳昱成
Chen, Yu-Cheng
關鍵詞 均數-變異數模型
風險趨避
強化學習
近端策略優化
Mean-Variance model
Risk Aversion
Reinforcement Learning
Proximal Policy Optimization
日期 2021
上傳時間 4-Aug-2021 14:51:05 (UTC+8)
摘要 Markowitz(1952)提出現代投資組合理論,透過均數-變異數模型(Mean-Variance Model) 為投資人進行資產配置,並設風險趨避參數調整報酬率和風險之間的比例,但在實務中,此風險趨避參數難以動態調整。本研究使用強化學習(Reinforcement Learning) 中的近端策略優化(Proximal Policy Optimization,PPO),依據不同市場變化,動態調整每一天風險趨避參數,當市場情況好時,投資人偏好承擔較高風險,獲得更高報酬,當市場情況壞時,投資人風險偏好趨於保守。本研究以台灣 50 指數當作整體市場走勢,比較強化學習輸入過去不同時間週期資訊之結果,研究結果顯示,不論輸入時間週期長短,強化學習績效皆能贏過固定風險趨避參數下均數-變異數模型,說明利用強化學習,能解決實務上風險趨避參數難以動態調整之問題。
Markowitz (1952) proposed Modern Portfolio Theory, which used the Mean-Variance Model to allocate assets for investors, and set the risk aversion parameter to adjust the ratio between return and the risk. But in practice, this risk aversion parameter is difficult to adjust dynamically. In our paper, we use Proximal Policy Optimization in reinforcement learning to dynamically adjust daily risk aversion parameters according to different market changes. In a bull market, investors prefer to take higher risks and get higher returns. On the other hand, in a bear market, investors` risk appetite tends to be conservative. This study uses the Taiwan 50 Index as the overall market trend, and compares the results of inputting different time periods of information into the model. The results show that regardless of the length of the input time period, the performance of the model can outperform the mean-variance model under fixed risk aversion parameters. Explain that the use of reinforcement learning can solve the problem of difficulty in dynamic adjustment of risk aversion parameters in practice.
參考文獻 劉上瑋 (2017)。深度增強學習在動態資產配置上之應用 : 以美國 ETF 為例。國立政治大學金融研究所碩士論文。
Basak, S., & Chabakauri, G. (2010). Dynamic mean-variance asset allocation. The Review of Financial Studies, 23(8), 2970-3016.
Björk, T., Murgoci, A., & Zhou, X. Y. (2014). Mean–variance portfolio optimization with state‐dependent risk aversion. Mathematical Finance: An International Journal of Mathematics, Statistics and Financial Economics, 24(1), 1-24.
Díaz, A., & Esparcia, C. (2019). Assessing risk aversion from the investor’s point of view. Frontiers in psychology, 10, 1490.
Gold, C. (2003, March). FX trading via recurrent reinforcement learning. In 2003 IEEE International Conference on Computational Intelligence for Financial Engineering, 2003. Proceedings. (pp. 363-370). IEEE.
Jiang, Z., & Liang, J. (2017, September). Cryptocurrency portfolio management with deep reinforcement learning. In 2017 Intelligent Systems Conference (IntelliSys) (pp. 905-913). IEEE.
Li, Y., & Li, Z. (2013). Optimal time-consistent investment and reinsurance strategies for mean–variance insurers with state dependent risk aversion. Insurance: Mathematics and Economics, 53(1), 86-97.
Markowitz, H. (1959). Portfolio selection. Journal of Finance, 7, 77–98.
Moody, J., & Saffell, M. (2001). Learning to trade via direct reinforcement. IEEE Transactions on Neural Networks, 12(4), 875-889.
Neuneier, R. (1998). Enhancing Q-learning for optimal asset allocation. Advances in neural information processing systems (pp. 936-942).
Rosenberg, J. V., & Engle, R. F. (2002). Empirical pricing kernels. Journal of Financial Economics, 64(3), 341-372.
Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347.
Zhang, Y., Wu, Y., Li, S., & Wiwatanapataphee, B. (2017). Mean-variance asset liability management with state-dependent risk aversion. North American Actuarial Journal, 21(1), 87-106.
Zhang, Y., Zhao, P., Li, B., Wu, Q., Huang, J., & Tan, M. (2020). Cost-sensitive portfolio selection via deep reinforcement learning. IEEE Transactions on Knowledge and Data Engineering.
描述 碩士
國立政治大學
金融學系
108352019
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0108352019
資料類型 thesis
dc.contributor.advisor 林士貴zh_TW
dc.contributor.author (Authors) 陳昱成zh_TW
dc.contributor.author (Authors) Chen, Yu-Chengen_US
dc.creator (作者) 陳昱成zh_TW
dc.creator (作者) Chen, Yu-Chengen_US
dc.date (日期) 2021en_US
dc.date.accessioned 4-Aug-2021 14:51:05 (UTC+8)-
dc.date.available 4-Aug-2021 14:51:05 (UTC+8)-
dc.date.issued (上傳時間) 4-Aug-2021 14:51:05 (UTC+8)-
dc.identifier (Other Identifiers) G0108352019en_US
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/136360-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 金融學系zh_TW
dc.description (描述) 108352019zh_TW
dc.description.abstract (摘要) Markowitz(1952)提出現代投資組合理論,透過均數-變異數模型(Mean-Variance Model) 為投資人進行資產配置,並設風險趨避參數調整報酬率和風險之間的比例,但在實務中,此風險趨避參數難以動態調整。本研究使用強化學習(Reinforcement Learning) 中的近端策略優化(Proximal Policy Optimization,PPO),依據不同市場變化,動態調整每一天風險趨避參數,當市場情況好時,投資人偏好承擔較高風險,獲得更高報酬,當市場情況壞時,投資人風險偏好趨於保守。本研究以台灣 50 指數當作整體市場走勢,比較強化學習輸入過去不同時間週期資訊之結果,研究結果顯示,不論輸入時間週期長短,強化學習績效皆能贏過固定風險趨避參數下均數-變異數模型,說明利用強化學習,能解決實務上風險趨避參數難以動態調整之問題。zh_TW
dc.description.abstract (摘要) Markowitz (1952) proposed Modern Portfolio Theory, which used the Mean-Variance Model to allocate assets for investors, and set the risk aversion parameter to adjust the ratio between return and the risk. But in practice, this risk aversion parameter is difficult to adjust dynamically. In our paper, we use Proximal Policy Optimization in reinforcement learning to dynamically adjust daily risk aversion parameters according to different market changes. In a bull market, investors prefer to take higher risks and get higher returns. On the other hand, in a bear market, investors` risk appetite tends to be conservative. This study uses the Taiwan 50 Index as the overall market trend, and compares the results of inputting different time periods of information into the model. The results show that regardless of the length of the input time period, the performance of the model can outperform the mean-variance model under fixed risk aversion parameters. Explain that the use of reinforcement learning can solve the problem of difficulty in dynamic adjustment of risk aversion parameters in practice.en_US
dc.description.tableofcontents 第一章 緒論 1
第一節 研究動機 1
第二節 研究目的 2
第二章 文獻回顧 4
第一節 投資組合理論 4
第二節 投資人風險偏好 4
第三節 強化學習之投資領域應用 5
第三章 研究方法 8
第一節 馬可維茲投資組合理論 8
第二節 強化學習 8
3.2.1 基本架構 9
3.2.2 策略梯度 9
3.2.3 近端策略優化(PPO) 12
3.2.4 強化學習動態風險趨避之應用 14
第三節 績效衡量指標 16
第四章 實證分析 17
第一節 資料來源與前處理 17
第二節 模型設定 18
第三節 實證結果 20
4.3.1 基準投資組合分析 20
4.3.2 強化學習投資組合績效分析 22
4.3.3 風險趨避程度與股價景氣之關係 28
第五章 結論與未來展望 30
參考文獻 32
zh_TW
dc.format.extent 2223290 bytes-
dc.format.mimetype application/pdf-
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0108352019en_US
dc.subject (關鍵詞) 均數-變異數模型zh_TW
dc.subject (關鍵詞) 風險趨避zh_TW
dc.subject (關鍵詞) 強化學習zh_TW
dc.subject (關鍵詞) 近端策略優化zh_TW
dc.subject (關鍵詞) Mean-Variance modelen_US
dc.subject (關鍵詞) Risk Aversionen_US
dc.subject (關鍵詞) Reinforcement Learningen_US
dc.subject (關鍵詞) Proximal Policy Optimizationen_US
dc.title (題名) 強化學習下動態調整風險偏好之投資組合配置:以台灣50指數為例zh_TW
dc.title (題名) Portfolio Allocation with Dynamic Risk Aversion via Reinforcement Learning: Evidence from Taiwan 50 Indexen_US
dc.type (資料類型) thesisen_US
dc.relation.reference (參考文獻) 劉上瑋 (2017)。深度增強學習在動態資產配置上之應用 : 以美國 ETF 為例。國立政治大學金融研究所碩士論文。
Basak, S., & Chabakauri, G. (2010). Dynamic mean-variance asset allocation. The Review of Financial Studies, 23(8), 2970-3016.
Björk, T., Murgoci, A., & Zhou, X. Y. (2014). Mean–variance portfolio optimization with state‐dependent risk aversion. Mathematical Finance: An International Journal of Mathematics, Statistics and Financial Economics, 24(1), 1-24.
Díaz, A., & Esparcia, C. (2019). Assessing risk aversion from the investor’s point of view. Frontiers in psychology, 10, 1490.
Gold, C. (2003, March). FX trading via recurrent reinforcement learning. In 2003 IEEE International Conference on Computational Intelligence for Financial Engineering, 2003. Proceedings. (pp. 363-370). IEEE.
Jiang, Z., & Liang, J. (2017, September). Cryptocurrency portfolio management with deep reinforcement learning. In 2017 Intelligent Systems Conference (IntelliSys) (pp. 905-913). IEEE.
Li, Y., & Li, Z. (2013). Optimal time-consistent investment and reinsurance strategies for mean–variance insurers with state dependent risk aversion. Insurance: Mathematics and Economics, 53(1), 86-97.
Markowitz, H. (1959). Portfolio selection. Journal of Finance, 7, 77–98.
Moody, J., & Saffell, M. (2001). Learning to trade via direct reinforcement. IEEE Transactions on Neural Networks, 12(4), 875-889.
Neuneier, R. (1998). Enhancing Q-learning for optimal asset allocation. Advances in neural information processing systems (pp. 936-942).
Rosenberg, J. V., & Engle, R. F. (2002). Empirical pricing kernels. Journal of Financial Economics, 64(3), 341-372.
Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347.
Zhang, Y., Wu, Y., Li, S., & Wiwatanapataphee, B. (2017). Mean-variance asset liability management with state-dependent risk aversion. North American Actuarial Journal, 21(1), 87-106.
Zhang, Y., Zhao, P., Li, B., Wu, Q., Huang, J., & Tan, M. (2020). Cost-sensitive portfolio selection via deep reinforcement learning. IEEE Transactions on Knowledge and Data Engineering.
zh_TW
dc.identifier.doi (DOI) 10.6814/NCCU202100814en_US