Publications-Theses
Article View/Open
Publication Export
-
題名 機器學習因子擇時模型結合Black-Litterman模型之投資組合建構
Portfolio Construction with Machine Learning Factor Timing and Black-Litterman Models作者 林生華
Lin, Sheng-Hua貢獻者 林靖庭
林生華
Lin, Sheng-Hua關鍵詞 因子投資
因子擇時
資產配置模型
隨機森林模型
橫斷面因子模型
Black-Litterman模型
五分位數投資組合策略
Factor investing
factor timing
asset allocation model
random forest model
cross-sectional factor model
Black-Litterman model
quintile portfolios日期 2020 上傳時間 3-Aug-2020 17:38:35 (UTC+8) 摘要 本研究融合因子投資、因子擇時、Black-Litterman資產配置模型等市場主流投資想法，以台灣上市股票作為資產池，建構投資策略，目標是建構穩健的投資組合，動態篩選有效因子，將有效因子融入權重優化過程，使得最終的資產配置權重能同時反映個股的優劣以及個股間相關性，動態配置資產。本研究之目的及研究成果，以下分述之 :•探討機器學習結合因子擇時模型之有效性樣本外期間，因子擇時模型準確率約為55%，當因子本身對於下期報酬有顯著影響力時，準確率更高。動量因子在樣本期間對於下期報酬不具影響力，然而其因子擇時模型則有60%以上的準確率，代表模型可以預測動量因子的有效性，具有擇時能力。•確認以橫斷面因子模型作為Black-Litterman之量化投資人觀點的可行性以Long-Short五分位數投資組合策略，分析分析有效合成因子之有效性，策略績效表現顯示，經因子擇時模型之有效合成因子其策略勝率高達74%，夏普比率為1.31。•研究結合因子擇時、量化投資人觀點、Black-Litterman權重配置而形成的投資策略之績效表現。考慮交易稅負，極大化夏普比率形成的投資組合，夏普比率為0.8，高於未經因子擇時模型之投資組合的夏普比率約1.78倍，統計上顯著異於大盤報酬，同時有較低的最大回撤比率。
In this study, we take the stocks listed on TSE as assets pool and construct a robust portfolio strategy with novel investment ideas, including factor investing, factor timing and Black-Litterman model. With this strategy, we can dynamically detect the efficient factors and composite these factors into single index to identify future performance of a stock. Also, by combining this index and portfolio optimizer, the weight dynamically changes due to this index and the correlations structure between stocks.The purposes and results of the study are listed below :•Show the efficacy of machine learning factor timing model.The averaging accuracy of factor timing models is about 0.55. The result also shows the fact that accuracy of factor model is positive correlative with degree of a factor’s efficiency.•Check the feasibility of quantitative investors’ view of Black-Litterman derived from cross-sectional factor model.We analyze efficacy of the efficient composite factor through quintile portfolio. The win rate of long-short strategy is 0.74, higher than benchmark. The Sharpe ratio is around 1.31 and beats the benchmark.•Show the performance of portfolio strategyThe Sharpe ratio of maximum Sharpe ratios strategy hits 0.8, approximately 1.78 times that of benchmark. Also, the mean return of this strategy statistically significantly differs from TAIEX.參考文獻 1. Black, F., & Litterman, R. (1990). Asset allocation: combining investor views with market equilibrium. Goldman Sachs Fixed Income Research, 115.2. Breiman, L., Friedman, J., Olshen, R. and Stone, C. (1984). Classification and Regression Trees. Wadsworth.3. Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.4. Connor, G. (1995). The three types of factor models: A comparison of their explanatory power. Financial Analysts Journal, 51(3), 42-46.5. Fama, E. F., & French, K. R. (1998). The Cross-Section of Expected StockReturns. Journal of Finance, 47(2), 427-465.6. Fama, E. F., & French, K. R. (1998). Value versus growth: The international evidence. Journal of finance, 53(6), 1975-1999.7. Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of financial economics, 116(1), 1-22.8. Figelman, I. (2017). Black–Litterman with a Factor Structure Applied to Multi-Asset Portfolios. The Journal of Portfolio Management, 44(2), 136-155.9. Frost, P. A., & Savarino, J. E. (1988). For better performance: Constrain portfolio weights. Journal of Portfolio Management, 15(1), 29.10. Grinold, R. C., and Kahn, R. N. (2000). Active Portfolio Management. NewYork: McGraw-Hill.11. Ho, T. K. (1998). The random subspace method for constructing decision forests. IEEE transactions on pattern analysis and machine intelligence, 20(8), 832-844.12. Hodges, P., Hogan, K., Peterson, J. R., & Ang, A. (2017). Factor timing with cross-sectional and time-series predictors. The Journal of Portfolio Management, 44(1), 30-43.13. Idzorek, T. (2007). A step-by-step guide to the Black-Litterman model: Incorporating user-specified confidence levels. In Forecasting expected returns in the financial markets (pp. 17-38). Academic Press.14. Lintner, J. (1965). The Valuation of Risk Assets and the Selection of RiskyInvestments in Stock Portfolios and Capital Budgets, The Review of Economics and Statistics, 47(1), 13-37.15. Markowitz, H. (1952). Portfolio selection, Journal of Finance, 7(1), 77-91.16. Miller, K. L., Li, H., Zhou, T. G., & Giamouridis, D. (2015). A risk-oriented model for factor timing decisions. The Journal of Portfolio Management, 41(3), 46-58.17. Mossin, J. (1966). Equilibrium in a Capital Asset Market, Econometrica, 34(4), 768-783.18. Rosenberg, B. (1974). Extra-market components of covariance in security returns. Journal of Financial and quantitative analysis, 263-274.19. Ross , S. A. (1976). The Arbitrage Theory of Capital Asset Pricing. Journal ofEconomic Theory, 13(3), 341-360.20. Sharpe, W. F. (1964), Capital Asset Price: A Theory of Market Equilibrium under Conditions of Risk, Journal of Finance, 19(3), 425-442.21. Theil, H. (1971). Principles of Econometrics. New York: Wiley and Sons.22. Theil, H. (1978). Introduction to Econometrics. New Jersey: Prentice-Hall, Inc. 描述 碩士
國立政治大學
金融學系
107352022資料來源 http://thesis.lib.nccu.edu.tw/record/#G0107352022 資料類型 thesis dc.contributor.advisor 林靖庭 zh_TW dc.contributor.author (Authors) 林生華 zh_TW dc.contributor.author (Authors) Lin, Sheng-Hua en_US dc.creator (作者) 林生華 zh_TW dc.creator (作者) Lin, Sheng-Hua en_US dc.date (日期) 2020 en_US dc.date.accessioned 3-Aug-2020 17:38:35 (UTC+8) - dc.date.available 3-Aug-2020 17:38:35 (UTC+8) - dc.date.issued (上傳時間) 3-Aug-2020 17:38:35 (UTC+8) - dc.identifier (Other Identifiers) G0107352022 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/130992 - 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) 107352022 zh_TW dc.description.abstract (摘要) 本研究融合因子投資、因子擇時、Black-Litterman資產配置模型等市場主流投資想法，以台灣上市股票作為資產池，建構投資策略，目標是建構穩健的投資組合，動態篩選有效因子，將有效因子融入權重優化過程，使得最終的資產配置權重能同時反映個股的優劣以及個股間相關性，動態配置資產。本研究之目的及研究成果，以下分述之 :•探討機器學習結合因子擇時模型之有效性樣本外期間，因子擇時模型準確率約為55%，當因子本身對於下期報酬有顯著影響力時，準確率更高。動量因子在樣本期間對於下期報酬不具影響力，然而其因子擇時模型則有60%以上的準確率，代表模型可以預測動量因子的有效性，具有擇時能力。•確認以橫斷面因子模型作為Black-Litterman之量化投資人觀點的可行性以Long-Short五分位數投資組合策略，分析分析有效合成因子之有效性，策略績效表現顯示，經因子擇時模型之有效合成因子其策略勝率高達74%，夏普比率為1.31。•研究結合因子擇時、量化投資人觀點、Black-Litterman權重配置而形成的投資策略之績效表現。考慮交易稅負，極大化夏普比率形成的投資組合，夏普比率為0.8，高於未經因子擇時模型之投資組合的夏普比率約1.78倍，統計上顯著異於大盤報酬，同時有較低的最大回撤比率。 zh_TW dc.description.abstract (摘要) In this study, we take the stocks listed on TSE as assets pool and construct a robust portfolio strategy with novel investment ideas, including factor investing, factor timing and Black-Litterman model. With this strategy, we can dynamically detect the efficient factors and composite these factors into single index to identify future performance of a stock. Also, by combining this index and portfolio optimizer, the weight dynamically changes due to this index and the correlations structure between stocks.The purposes and results of the study are listed below :•Show the efficacy of machine learning factor timing model.The averaging accuracy of factor timing models is about 0.55. The result also shows the fact that accuracy of factor model is positive correlative with degree of a factor’s efficiency.•Check the feasibility of quantitative investors’ view of Black-Litterman derived from cross-sectional factor model.We analyze efficacy of the efficient composite factor through quintile portfolio. The win rate of long-short strategy is 0.74, higher than benchmark. The Sharpe ratio is around 1.31 and beats the benchmark.•Show the performance of portfolio strategyThe Sharpe ratio of maximum Sharpe ratios strategy hits 0.8, approximately 1.78 times that of benchmark. Also, the mean return of this strategy statistically significantly differs from TAIEX. en_US dc.description.tableofcontents 中文摘要 IABSTRACT II表次 IV圖次 V第一章 緒論 1第一節 研究背景及動機 1第二節 研究目的 2第三節 研究架構 2第二章 文獻探討 4第一節 現代投資組合理論 4第二節 資產定價模型 4第三節 多因子模型 5第四節 因子擇時模型 5第三章 研究方法 7第一節 研究流程 7第二節 資料說明 10第三節 隨機森林因子擇時模型 12第四節 橫斷面因子模型 22第五節 BLACK-LITTERMAN模型 24第六節 投資組合績效評估 34第四章 實證結果 36第一節 隨機森林因子擇時模型實證結果 36第二節 量化投資人觀點之預期報酬分析 40第三節 投資組合績效回測 41第五章 結論及建議 45第一節 結論 45第二節 建議及後續研究 45參考文獻 47附錄一、因子定義及計算方式 49附錄二、輸入變數敘述統計表-因子投組橫斷面特徵 53附錄三、輸入變數敘述統計表-因子投組風險集中度及總體經濟及市場風險曝險 61附錄四、投資組合權重比較圖表 69 zh_TW dc.format.extent 3021459 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0107352022 en_US dc.subject (關鍵詞) 因子投資 zh_TW dc.subject (關鍵詞) 因子擇時 zh_TW dc.subject (關鍵詞) 資產配置模型 zh_TW dc.subject (關鍵詞) 隨機森林模型 zh_TW dc.subject (關鍵詞) 橫斷面因子模型 zh_TW dc.subject (關鍵詞) Black-Litterman模型 zh_TW dc.subject (關鍵詞) 五分位數投資組合策略 zh_TW dc.subject (關鍵詞) Factor investing en_US dc.subject (關鍵詞) factor timing en_US dc.subject (關鍵詞) asset allocation model en_US dc.subject (關鍵詞) random forest model en_US dc.subject (關鍵詞) cross-sectional factor model en_US dc.subject (關鍵詞) Black-Litterman model en_US dc.subject (關鍵詞) quintile portfolios en_US dc.title (題名) 機器學習因子擇時模型結合Black-Litterman模型之投資組合建構 zh_TW dc.title (題名) Portfolio Construction with Machine Learning Factor Timing and Black-Litterman Models en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) 1. Black, F., & Litterman, R. (1990). Asset allocation: combining investor views with market equilibrium. Goldman Sachs Fixed Income Research, 115.2. Breiman, L., Friedman, J., Olshen, R. and Stone, C. (1984). Classification and Regression Trees. Wadsworth.3. Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.4. Connor, G. (1995). The three types of factor models: A comparison of their explanatory power. Financial Analysts Journal, 51(3), 42-46.5. Fama, E. F., & French, K. R. (1998). The Cross-Section of Expected StockReturns. Journal of Finance, 47(2), 427-465.6. Fama, E. F., & French, K. R. (1998). Value versus growth: The international evidence. Journal of finance, 53(6), 1975-1999.7. Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of financial economics, 116(1), 1-22.8. Figelman, I. (2017). Black–Litterman with a Factor Structure Applied to Multi-Asset Portfolios. The Journal of Portfolio Management, 44(2), 136-155.9. Frost, P. A., & Savarino, J. E. (1988). For better performance: Constrain portfolio weights. Journal of Portfolio Management, 15(1), 29.10. Grinold, R. C., and Kahn, R. N. (2000). Active Portfolio Management. NewYork: McGraw-Hill.11. Ho, T. K. (1998). The random subspace method for constructing decision forests. IEEE transactions on pattern analysis and machine intelligence, 20(8), 832-844.12. Hodges, P., Hogan, K., Peterson, J. R., & Ang, A. (2017). Factor timing with cross-sectional and time-series predictors. The Journal of Portfolio Management, 44(1), 30-43.13. Idzorek, T. (2007). A step-by-step guide to the Black-Litterman model: Incorporating user-specified confidence levels. In Forecasting expected returns in the financial markets (pp. 17-38). Academic Press.14. Lintner, J. (1965). The Valuation of Risk Assets and the Selection of RiskyInvestments in Stock Portfolios and Capital Budgets, The Review of Economics and Statistics, 47(1), 13-37.15. Markowitz, H. (1952). Portfolio selection, Journal of Finance, 7(1), 77-91.16. Miller, K. L., Li, H., Zhou, T. G., & Giamouridis, D. (2015). A risk-oriented model for factor timing decisions. The Journal of Portfolio Management, 41(3), 46-58.17. Mossin, J. (1966). Equilibrium in a Capital Asset Market, Econometrica, 34(4), 768-783.18. Rosenberg, B. (1974). Extra-market components of covariance in security returns. Journal of Financial and quantitative analysis, 263-274.19. Ross , S. A. (1976). The Arbitrage Theory of Capital Asset Pricing. Journal ofEconomic Theory, 13(3), 341-360.20. Sharpe, W. F. (1964), Capital Asset Price: A Theory of Market Equilibrium under Conditions of Risk, Journal of Finance, 19(3), 425-442.21. Theil, H. (1971). Principles of Econometrics. New York: Wiley and Sons.22. Theil, H. (1978). Introduction to Econometrics. New Jersey: Prentice-Hall, Inc. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU202001165 en_US
