1. NCCU Academic Hub
  2. Publications
  3. College of Commerce
  4. Theses
  5. Item

Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 以機器學習探討動能現象 -- 以台灣股市為例
Exploring Momentum Phenomena in Taiwan Stock Market through Machine Learning
作者 章翔軒
Chang, Hsiang-Hsuan
貢獻者 羅秉政
Kendro Vincent
章翔軒
Chang,Hsiang-Hsuan
關鍵詞 橫斷面動能
時間序列動能
機器學習
投資組合
Cross-Sectional Momentum
Time Series Momentum
Machine Learning
Portfolio
日期 2025
上傳時間 4-Aug-2025 14:31:21 (UTC+8)
摘要 在股市動能現象的研究中，主要分為橫斷面動能（Cross-Sectional Momentum）與時間序列動能（Time Series Momentum）。前者著重於資產在同期相較於其他資產的報酬率如果具有相對優勢，此趨勢預期將會持續；後者則關注某個資產的報酬率是否為正，若資產的報酬率大於0，則預期該趨勢將會延續。本研究以台灣股票為樣本，透過機器學習預測的方式探討兩種動能現象，透過羅吉斯迴歸、決策樹集成式學習(XGBoost、隨機森林)與類神經網路進行預測；接者將預測結果以 Goyal and Jegadeesh (2018) 的加權方式形成不同投資組合，比較各模型和加權方式之間的獲利表現差異。實證結果顯示，橫斷面動能做為被解釋變數在同模型預測的準確率高於時間序列動能，但整體預測上具有提升空間；橫斷面動能投資組合報酬多數高於時間序列動能投資組合報酬，但只有部分模型預測組建的橫斷面動能組合在報酬率上勝過純粹以報酬率建構的基準動能組合。
In the study of stock market momentum phenomena, two primary types are commonly examined: Cross-Sectional Momentum (CSM) and Time Series Momentum (TSM). CSM focuses on the relative performance of assets compared to one another during the same period—if an asset outperforms its peers, the trend is expected to continue. In contrast, TSM emphasizes the past return of a single asset—if an asset’s return is positive, it is expected that the trend will persist. This study uses Taiwan’s stock market as the empirical setting and applies machine learning techniques to investigate both types of momentum. Models used include logistic regression, tree-based ensemble methods (XGBoost and Random Forest), and neural networks. Based on the prediction results, investment portfolios are constructed following the weighting methodology proposed by Goyal and Jegadeesh (2018), and the profitability across models and weighting schemes is compared. The empirical findings show that using CSM as the target variable leads to higher prediction accuracy than TSM under the same model framework. However, overall predictive performance still has room for improvement. In terms of portfolio returns, CSM-based strategies generally outperform those based on TSM. Nonetheless, only a subset of CSM portfolios constructed from machine learning predictions achieved higher returns than the benchmark momentum portfolios built solely on raw return rankings.
參考文獻 1. Ballings, M., Van den Poel, D., Hespeels, N., & Gryp, R. (2015). Evaluating multiple classifiers for stock price direction prediction. Expert Systems with Applications, 42(20), 7046–7056. 2. Bui, D. G., Kong, D.-R., Lin, C.-Y., & Lin, T.-C. (2023). Momentum in machine learning: Evidence from the Taiwan stock market. Pacific-Basin Finance Journal, 82, 102178. 3. Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of Finance, 52(1), 57–82. 4. Chakrabarti, G. (2015). Time-series momentum trading strategies in the global stock market. Business Economics, 50, 80–90. 5. Chen, T., & Guestrin, C. (2016). Xgboost: A scalable tree boosting system. Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, 785–794. 6. Cujean, J., & Hasler, M. (2017). Why does return predictability concentrate in bad times? The Journal of Finance, 72(6), 2717–2758. 7. D’Souza, I., Srichanachaichok, V., Wang, G. J., & Yao, C. Y. (2016). The enduring effect of time-series momentum on stock returns over nearly 100-years. Asian Finance Association (AsianFA) 2016 Conference. 8. Daniel, K., & Moskowitz, T. J. (2016). Momentum crashes. Journal of Financial Economics, 122(2), 221–247. 9. Deboeck, G. J. (1994). Trading on the edge: neural, genetic, and fuzzy systems for chaotic financial markets (Vol. 39). John Wiley & Sons. 10. Fama, E. F. (1970). Efficient capital markets. Journal of Finance, 25(2), 383–417. 11. Fama, E. F., & French, K. R. (1988). Permanent and temporary components of stock prices. Journal of Political Economy, 96(2), 246–273. 12. ———. (1996). Multifactor explanations of asset pricing anomalies. The Journal of Finance, 51(1), 55–84. 13. ———. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1–22. 14. Fischer, T., & Krauss, C. (2018). Deep learning with long short-term memory networks for financial market predictions. European Journal of Operational Research, 270(2), 654–669. 15. Goyal, A., & Jegadeesh, N. (2018). Cross-sectional and time-series tests of return predictability: What is the difference? The Review of Financial Studies, 31(5), 1784–1824. 16. Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. The Review of Financial Studies, 33(5), 2223–2273. 17. He, X.-Z., & Li, K. (2015). Profitability of time series momentum. Journal of Banking & Finance, 53, 140–157. 18. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780. 19. Hong, H., & Stein, J. C. (1999). A unified theory of underreaction, momentum trading, and overreaction in asset markets. The Journal of Finance, 54(6), 2143–2184. 20. Huang, D., Li, J., Wang, L., & Zhou, G. (2020). Time series momentum: Is it there? Journal of Financial Economics, 135(3), 774–794. 21. Jegadeesh, N. (1990). Evidence of predictable behavior of security returns. The Journal of Finance, 45(3), 881–898. 22. Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. The Journal of Finance, 48(1), 65–91. 23. Kim, A. Y., Tse, Y., & Wald, J. K. (2016). Time series momentum and volatility scaling. Journal of Financial Markets, 30, 103–124. 24. Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv Preprint arXiv:1412.6980. 25. Krauss, C., Do, X. A., & Huck, N. (2017). Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500. European Journal of Operational Research, 259(2), 689–702. 26. Lehmann, B. N. (1990). Fads, martingales, and market efficiency. The Quarterly Journal of Economics, 105(1), 1–28. 27. Leung, M. T., Daouk, H., & Chen, A.-S. (2000). Forecasting stock indices: a comparison of classification and level estimation models. International Journal of Forecasting, 16(2), 173–190. 28. Lim, B. Y., Wang, J. G., & Yao, Y. (2018). Time-series momentum in nearly 100 years of stock returns. Journal of Banking & Finance, 97, 283–296. 29. Lim, B., Zohren, S., & Roberts, S. (2019). Enhancing time series momentum strategies using deep neural networks. arXiv Preprint arXiv:1904.04912. 30. Lo, A. W. (1991). Long-term memory in stock market prices. Econometrica: Journal of the Econometric Society, 1279–1313. 31. Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30. 32. Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012). Time series momentum. Journal of Financial Economics, 104(2), 228–250. 33. Pan, S., Long, S. C., Wang, Y., & Xie, Y. (2023). Nonlinear asset pricing in Chinese stock market: A deep learning approach. International Review of Financial Analysis, 87, 102627. 34. Pesaran, M. H., & Timmermann, A. (1992). A simple nonparametric test of predictive performance. Journal of Business & Economic Statistics, 10(4), 461–465. 35. Qin, Y., Pan, G., & Bai, M. (2020). Improving market timing of time series momentum in the Chinese stock market. Applied Economics, 52(43), 4711–4725. 36. Takeuchi, L., & Lee, Y.-Y. A. (2013). Applying deep learning to enhance momentum trading strategies in stocks. In Technical Report. Stanford University Stanford, CA, USA. 37. Zakamulin, V., & Giner, J. (2022). Time series momentum in the US stock market: Empirical evidence and theoretical analysis. International Review of Financial Analysis, 82, 102173.
描述 碩士
國立政治大學
金融學系
110352031
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0110352031
資料類型 thesis
dc.contributor.advisor 羅秉政zh_TW
dc.contributor.advisor Kendro Vincenten_US
dc.contributor.author (Authors) 章翔軒zh_TW
dc.contributor.author (Authors) Chang,Hsiang-Hsuanen_US
dc.creator (作者) 章翔軒zh_TW
dc.creator (作者) Chang, Hsiang-Hsuanen_US
dc.date (日期) 2025en_US
dc.date.accessioned 4-Aug-2025 14:31:21 (UTC+8)-
dc.date.available 4-Aug-2025 14:31:21 (UTC+8)-
dc.date.issued (上傳時間) 4-Aug-2025 14:31:21 (UTC+8)-
dc.identifier (Other Identifiers) G0110352031en_US
dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/158584-
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) 110352031zh_TW
dc.description.abstract (摘要) 在股市動能現象的研究中，主要分為橫斷面動能（Cross-Sectional Momentum）與時間序列動能（Time Series Momentum）。前者著重於資產在同期相較於其他資產的報酬率如果具有相對優勢，此趨勢預期將會持續；後者則關注某個資產的報酬率是否為正，若資產的報酬率大於0，則預期該趨勢將會延續。本研究以台灣股票為樣本，透過機器學習預測的方式探討兩種動能現象，透過羅吉斯迴歸、決策樹集成式學習(XGBoost、隨機森林)與類神經網路進行預測；接者將預測結果以 Goyal and Jegadeesh (2018) 的加權方式形成不同投資組合，比較各模型和加權方式之間的獲利表現差異。實證結果顯示，橫斷面動能做為被解釋變數在同模型預測的準確率高於時間序列動能，但整體預測上具有提升空間；橫斷面動能投資組合報酬多數高於時間序列動能投資組合報酬，但只有部分模型預測組建的橫斷面動能組合在報酬率上勝過純粹以報酬率建構的基準動能組合。zh_TW
dc.description.abstract (摘要) In the study of stock market momentum phenomena, two primary types are commonly examined: Cross-Sectional Momentum (CSM) and Time Series Momentum (TSM). CSM focuses on the relative performance of assets compared to one another during the same period—if an asset outperforms its peers, the trend is expected to continue. In contrast, TSM emphasizes the past return of a single asset—if an asset’s return is positive, it is expected that the trend will persist. This study uses Taiwan’s stock market as the empirical setting and applies machine learning techniques to investigate both types of momentum. Models used include logistic regression, tree-based ensemble methods (XGBoost and Random Forest), and neural networks. Based on the prediction results, investment portfolios are constructed following the weighting methodology proposed by Goyal and Jegadeesh (2018), and the profitability across models and weighting schemes is compared. The empirical findings show that using CSM as the target variable leads to higher prediction accuracy than TSM under the same model framework. However, overall predictive performance still has room for improvement. In terms of portfolio returns, CSM-based strategies generally outperform those based on TSM. Nonetheless, only a subset of CSM portfolios constructed from machine learning predictions achieved higher returns than the benchmark momentum portfolios built solely on raw return rankings.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 第三節 研究架構 4 第二章 文獻回顧 5 第一節 時間序列動能相關文獻 5 第二節 機器學習之相關文獻 8 第三章 研究方法 12 第一節 羅吉斯迴歸 12 第二節 決策樹模型 13 一、 XGBoost 15 二、 隨機森林 18 第四節 類神經網路 20 第五節 LSTM 24 第六節 模型訓練方法 26 第七節 模型預測能力指標 30 一、混淆矩陣與方向判定檢定 30 二、 特徵重要性分析指標 32 第八節 投資組合建構方法 34 第四章 實證分析 37 第一節 資料描述 37 第二節 模型預測能力分析 38 一、 預測表現 38 二、特徵重要性 43 第三節 動能組合獲利能力分析 46 一、組合報酬率 46 二、風險調整報酬率 56 第四節 外在因素探討 60 一、公司市值 60 二、市場狀態 66 第五章 結論與未來展望 69 第一節 研究結論 69 第二節 未來研究方向 70 參考文獻 71 附錄一 75 附錄二 76zh_TW
dc.format.extent 5379597 bytes-
dc.format.mimetype application/pdf-
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0110352031en_US
dc.subject (關鍵詞) 橫斷面動能zh_TW
dc.subject (關鍵詞) 時間序列動能zh_TW
dc.subject (關鍵詞) 機器學習zh_TW
dc.subject (關鍵詞) 投資組合zh_TW
dc.subject (關鍵詞) Cross-Sectional Momentumen_US
dc.subject (關鍵詞) Time Series Momentumen_US
dc.subject (關鍵詞) Machine Learningen_US
dc.subject (關鍵詞) Portfolioen_US
dc.title (題名) 以機器學習探討動能現象 -- 以台灣股市為例zh_TW
dc.title (題名) Exploring Momentum Phenomena in Taiwan Stock Market through Machine Learningen_US
dc.type (資料類型) thesisen_US
dc.relation.reference (參考文獻) 1. Ballings, M., Van den Poel, D., Hespeels, N., & Gryp, R. (2015). Evaluating multiple classifiers for stock price direction prediction. Expert Systems with Applications, 42(20), 7046–7056. 2. Bui, D. G., Kong, D.-R., Lin, C.-Y., & Lin, T.-C. (2023). Momentum in machine learning: Evidence from the Taiwan stock market. Pacific-Basin Finance Journal, 82, 102178. 3. Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of Finance, 52(1), 57–82. 4. Chakrabarti, G. (2015). Time-series momentum trading strategies in the global stock market. Business Economics, 50, 80–90. 5. Chen, T., & Guestrin, C. (2016). Xgboost: A scalable tree boosting system. Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, 785–794. 6. Cujean, J., & Hasler, M. (2017). Why does return predictability concentrate in bad times? The Journal of Finance, 72(6), 2717–2758. 7. D’Souza, I., Srichanachaichok, V., Wang, G. J., & Yao, C. Y. (2016). The enduring effect of time-series momentum on stock returns over nearly 100-years. Asian Finance Association (AsianFA) 2016 Conference. 8. Daniel, K., & Moskowitz, T. J. (2016). Momentum crashes. Journal of Financial Economics, 122(2), 221–247. 9. Deboeck, G. J. (1994). Trading on the edge: neural, genetic, and fuzzy systems for chaotic financial markets (Vol. 39). John Wiley & Sons. 10. Fama, E. F. (1970). Efficient capital markets. Journal of Finance, 25(2), 383–417. 11. Fama, E. F., & French, K. R. (1988). Permanent and temporary components of stock prices. Journal of Political Economy, 96(2), 246–273. 12. ———. (1996). Multifactor explanations of asset pricing anomalies. The Journal of Finance, 51(1), 55–84. 13. ———. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1–22. 14. Fischer, T., & Krauss, C. (2018). Deep learning with long short-term memory networks for financial market predictions. European Journal of Operational Research, 270(2), 654–669. 15. Goyal, A., & Jegadeesh, N. (2018). Cross-sectional and time-series tests of return predictability: What is the difference? The Review of Financial Studies, 31(5), 1784–1824. 16. Gu, S., Kelly, B., & Xiu, D. (2020). Empirical asset pricing via machine learning. The Review of Financial Studies, 33(5), 2223–2273. 17. He, X.-Z., & Li, K. (2015). Profitability of time series momentum. Journal of Banking & Finance, 53, 140–157. 18. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780. 19. Hong, H., & Stein, J. C. (1999). A unified theory of underreaction, momentum trading, and overreaction in asset markets. The Journal of Finance, 54(6), 2143–2184. 20. Huang, D., Li, J., Wang, L., & Zhou, G. (2020). Time series momentum: Is it there? Journal of Financial Economics, 135(3), 774–794. 21. Jegadeesh, N. (1990). Evidence of predictable behavior of security returns. The Journal of Finance, 45(3), 881–898. 22. Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers: Implications for stock market efficiency. The Journal of Finance, 48(1), 65–91. 23. Kim, A. Y., Tse, Y., & Wald, J. K. (2016). Time series momentum and volatility scaling. Journal of Financial Markets, 30, 103–124. 24. Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv Preprint arXiv:1412.6980. 25. Krauss, C., Do, X. A., & Huck, N. (2017). Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500. European Journal of Operational Research, 259(2), 689–702. 26. Lehmann, B. N. (1990). Fads, martingales, and market efficiency. The Quarterly Journal of Economics, 105(1), 1–28. 27. Leung, M. T., Daouk, H., & Chen, A.-S. (2000). Forecasting stock indices: a comparison of classification and level estimation models. International Journal of Forecasting, 16(2), 173–190. 28. Lim, B. Y., Wang, J. G., & Yao, Y. (2018). Time-series momentum in nearly 100 years of stock returns. Journal of Banking & Finance, 97, 283–296. 29. Lim, B., Zohren, S., & Roberts, S. (2019). Enhancing time series momentum strategies using deep neural networks. arXiv Preprint arXiv:1904.04912. 30. Lo, A. W. (1991). Long-term memory in stock market prices. Econometrica: Journal of the Econometric Society, 1279–1313. 31. Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30. 32. Moskowitz, T. J., Ooi, Y. H., & Pedersen, L. H. (2012). Time series momentum. Journal of Financial Economics, 104(2), 228–250. 33. Pan, S., Long, S. C., Wang, Y., & Xie, Y. (2023). Nonlinear asset pricing in Chinese stock market: A deep learning approach. International Review of Financial Analysis, 87, 102627. 34. Pesaran, M. H., & Timmermann, A. (1992). A simple nonparametric test of predictive performance. Journal of Business & Economic Statistics, 10(4), 461–465. 35. Qin, Y., Pan, G., & Bai, M. (2020). Improving market timing of time series momentum in the Chinese stock market. Applied Economics, 52(43), 4711–4725. 36. Takeuchi, L., & Lee, Y.-Y. A. (2013). Applying deep learning to enhance momentum trading strategies in stocks. In Technical Report. Stanford University Stanford, CA, USA. 37. Zakamulin, V., & Giner, J. (2022). Time series momentum in the US stock market: Empirical evidence and theoretical analysis. International Review of Financial Analysis, 82, 102173.zh_TW