學術產出-Theses
Article View/Open
Publication Export
-
題名 應用機器學習於外匯超額報酬預測
Applying Machine Learning for the Prediction of Foreign Exchange Return作者 朱珮錡
Chu, Pei-Chi貢獻者 林建秀
Lin, Chien-Hsiu
朱珮錡
Chu, Pei-Chi關鍵詞 外匯交易策略
機器學習
隨機森林
梯度提升樹
極限梯度提升樹
Forex strategy
Machine learning
Random Forest
Gradient Boosting Decision Tree
Extreme Gradient Boosting日期 2021 上傳時間 4-Aug-2021 14:51:45 (UTC+8) 摘要 影響匯率走勢的因素有央行利率決策、一國之經濟數據及當地政府治理、政治因素等，透過分析重要因素預測匯率未來變化，能讓企業使用較少成本做匯率避險、外匯投機者能從交易策略中賺取報酬。本研究將機器學習技術應用於外匯市場上，結合影響匯率走勢的重要因素及機器學習模型，嘗試對匯率變動方向作預測，以十六個國家的外匯資料及六十個因子為研究變數。本研究藉由隨機森林(Random Forest)、梯度提升樹(GBDT)及極限梯度提升樹(XGBoost)演算法預測匯率走勢，並比較各模型的預測準確度，以及統整出影響匯率走勢之重要因素，也將模型的預測結果建構出投資組合，並使用評量指標分析模型的績效表現。本研究也利用實證上有效的因子，如：利差因子、動能因子、價值因子、市場因子，將此四個因子同時納入機器學習模型及羅吉斯回歸中，比較四因子使用不同方式做預測的差異，也將預測結果建構成投資組合，與實證上有超額報酬的外匯交易策略做比較。實證結果顯示，機器學習模型在預測匯率變動方向上比羅吉斯回歸佳，顯示機器學習確實有分類預測能力；使用六十個因子的預測準確度優於使用四個因子，顯示較多因子能涵蓋更多資訊，可得到更好預測結果；三種機器學習模型的預測準確度差異不大。在模型之重要因子方面，三種機器學習模型中，政府效能都是相當重要的因子，顯示一國政府治理、政治局勢是否穩定是影響匯率重要因素。在投資組合績效表現上，四因子納入機器學習模型優於實證上有效的單一策略，顯示多因子機器學習模型建構出來的策略比單一因子更好且更穩定；另外，也發現GBDT與XGBoost的投資組合有較高累積報酬，表現優於隨機森林。
Forex rates can be predicted by analyzing factors, like interest rate decisions, economic data, and political situations. If companies can have better opinions about forex rates, they can use less cost for hedging or speculators can earn returns from some strategies. In this study, I used factors in machine learning models to predict forex rates.In this study, Random Forest, Gradient Boosting Decision Tree and Extreme Gradient Boosting are used to predict forex rates. Forecast accuracy are compared in three models and important factors are found out. Predictions are formed a portfolio and the performance are compared between models. I also use the factors of Carry, Market, Momentum and Value in machine learning models and Logistic Regression to compare predictions and performance of trading strategies between models.Results show machine learning models are better than logistic regression in predicting the direction of forex changes. Besides, the accuracy of sixty factors is better than four factors. However, the accuracy between three machine learning models is not much different. In the feature importance, I found factors related to governance are important in all machine learning models. In the portfolio performance, using four factors is better than a single strategy, showing that a portfolio constructed by multi-factor machine learning model is better than a single factor. Besides, the study found GBDT and XGBoost are better than Random Forest in performance.參考文獻 [1] 郭秀樺. (2018). 外匯報酬之利差、動能及價值交易策略成因分析. 國立政治大學碩士學位論文.[2] 林庭陞. (2020). 機器學習匯率定價投資組合. 國立政治大學碩士學位論文.[3] Acuña, E. & Rodriguez, C. (2004). The Treatment of Missing Values and its Effect on Classifier Accuracy. In: Banks D., McMorris F.R., Arabie P., Gaul W. (eds) Classification, Clustering, and Data Mining Applications. Studies in Classification, Data Analysis, and Knowledge Organisation. Springer, Berlin, Heidelberg.[4] Alsaleem, M.Y.A., & Hasoon, S.O. (2020). Comparison of dt& gbdt algorithms for predictive modeling of currency exchange rates. EUREKA: Physics and Engineering.[5] Asness, C.S., Moskowitz, T.J., & Pedersen, L.H. (2013). Value and Momentum Everywhere. The Journal of Finance, 68(3), 929-985.[6] Berg, K.A. & Mark, N.C. (2018). Measures of global uncertainty and carry-trade excess returns. Journal of International Money and Finance, 88, 212-227.[7] Booth, A., Gerding, E., & McGroarty, F. (2014). Automated trading with performance weighted random forests and seasonality. Expert Systems with Applications, 41(8), 3651-3661.[8] Bousbaa, Z., Chihab, M., Bencharef, O., & Zitiet, S. (2019). Algo-Trading Strategy for Intraweek Foreign Exchange Speculation Based on Random Forest and Probit Regression. Applied Computational Intelligence and Soft Computing.[9] Breiman, L. (2001). Random Forests. Machine Learning, 45, 5-32.[10] Brunnermeier, M.K., Nagel, S., & Pedersen, L.H. (2008). Carry Trades and Currency Crashes. NBER Macroeconomics Annual, 23.[11] Burnside, C., Eichenbaum, M., & Rebelo, S. (2011). Carry Trade and Momentum in Currency Markets. Annual Review of Financial Economics, 3, 511-535.[12] Chan, L.K.C., Jegadeesh, N., & Lakonishok, J. (1995). Evaluating the performance of value versus glamour stocks the impact of selection bias. Journal of Financial Economics, 38(3), 269-296.[13] Chen, C.N., & Lin, C.H. (2020). The sources of pricing factors underlying the cross-section of currency returns. The Quarterly Review of Economics and Finance, 77, 250-265.[14] Chen, T. & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. KDD `16: The 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco California, USA, 785-794.[15] Dey, S., Kumar, Y., Saha, S. & Basak, S. (2016). Forecasting to Classification: Predicting the direction of stock market price using Xtreme Gradient Boosting.[16] Du, W., Tepper, A., & Verdelhan, A. (2018). Deviations from covered interest rate parity. The Journal of Finance, 73(3), 915-957.[17] Fama, E.F., & French, K.R. (1992). The Cross-Section of Expected Stock Returns. The Journal of Finance, 47(2), 427-465.[18] Ferrari, M., Kearns, J. & Schrimpf, A. (2021). Monetary policy’s rising FX impact in the era of ultra-low rates. Journal of Banking & Finance, 129.[19] Filardo, A.J., Mohanty, M.S. & Moreno, R. (2012). Central Bank and Government Debt Management: Issues for Monetary Policy. BIS Paper No. 67d.[20] Friedman, J.H. (2001). Greedy Function Approximation: A Gradient Boosting Machine. The Annals of Statistics, 29(5), 1189-1232.[21] Galeshchuk, S. & Mukherjee, S. (2017). Deep networks for predicting direction of change in foreign exchange rates. Intelligent Systems in Accounting Finance & Management, 24(3).[22] Gu, S., Kelly, B. & Xiu, D. (2020). Empirical Asset Pricing via Machine Learning. The Review of Financial Studies, 33(5), 2223-2273.[23] Islam, S.F.N., Sholahuddin, A. & Abdullah, A.S. (2021). Extreme gradient boosting (XGBoost) method in making forecasting application and analysis of USD exchange rates against rupiah. Tenth International Conference and Workshop on High Dimensional Data Analysis (ICW-HDDA-X) 12-15 October 2020 in Sanur-Bali, Indonesia, 1722.[24] 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.[25] Kroencke, T.A., Schindler, F., & Schrimpf, A. (2014). International diversification benefits with foreign exchange investment styles. Review of Finance, 18(5), 1847- 1883.[26] Lintner, J. (1965). Security Prices, Risk, and Maximal Gains from Diversification. The Journal of Finance, 20(4), 587-615.[27] Liu, H. (2020). Stock Selection Strategy Based on Support Vector Machine and eXtreme Gradient Boosting Methods. 2020 the 4th International Conference on Big Data Research, 36-39.[28] Lukas, M., & Mark, P.T. (2007). The Obstinate Passion of Foreign Exchange Professionals: Technical Analysis. Journal of Economic. Literature, 45(4), 936-972.[29] Lustig, H., Roussanov, N., & Verdelhan, A. (2011). Common Risk Factor in Currency Markets. Review of Financial Studies. 24(11), 3731-3777.[30] Maragoudakis, M. & Serpanos, D. (2010). Towards Stock Market Data Mining Using Enriched Random Forests from Textual Resources and Technical Indicators.[31] Menkhoff, L., Sarno, L., Shmeling, M., & Schrimpf, A. (2012). Currency Momentum Strategies. Journal of Financial Economics, 106(3), 660-684.[32] Okunev, J., & White, D. (2003). Do Momentum-Based Strategies Still Work in Foreign Currency Markets? The Journal of Financial and Quantitative Analysis, 38(2), 425-447.[33] Qian, B. & Rasheed, K. (2010). Foreign exchange market prediction with multiple classifiers. Journal of Forecasting, 29(3), 271-284.[34] Qin, Q., Wang, Q.G., Li, J. & Ge, S.S. (2013). Linear and Nonlinear Trading Models with Gradient Boosted Random Forests and Application to Singapore Stock Market. Journal of Intelligent Learning Systems and Applications, 5(1).[35] Raza, A. (2015). Are Value Strategies Profitable in the Foreign Exchange Market? 28th Australasian Finance and Banking Conference[36] Schut, F.G.B., van Rijn, J.N., & Hoos, H.H. (2015). Towards Automated Technical Analysis for Foreign Exchange Data.[37] Sharpe, W.F. (1964). Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk. The Journal of Finance, 19(3), 425-442. 描述 碩士
國立政治大學
金融學系
108352023資料來源 http://thesis.lib.nccu.edu.tw/record/#G0108352023 資料類型 thesis dc.contributor.advisor 林建秀 zh_TW dc.contributor.advisor Lin, Chien-Hsiu en_US dc.contributor.author (Authors) 朱珮錡 zh_TW dc.contributor.author (Authors) Chu, Pei-Chi en_US dc.creator (作者) 朱珮錡 zh_TW dc.creator (作者) Chu, Pei-Chi en_US dc.date (日期) 2021 en_US dc.date.accessioned 4-Aug-2021 14:51:45 (UTC+8) - dc.date.available 4-Aug-2021 14:51:45 (UTC+8) - dc.date.issued (上傳時間) 4-Aug-2021 14:51:45 (UTC+8) - dc.identifier (Other Identifiers) G0108352023 en_US dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/136363 - 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) 108352023 zh_TW dc.description.abstract (摘要) 影響匯率走勢的因素有央行利率決策、一國之經濟數據及當地政府治理、政治因素等，透過分析重要因素預測匯率未來變化，能讓企業使用較少成本做匯率避險、外匯投機者能從交易策略中賺取報酬。本研究將機器學習技術應用於外匯市場上，結合影響匯率走勢的重要因素及機器學習模型，嘗試對匯率變動方向作預測，以十六個國家的外匯資料及六十個因子為研究變數。本研究藉由隨機森林(Random Forest)、梯度提升樹(GBDT)及極限梯度提升樹(XGBoost)演算法預測匯率走勢，並比較各模型的預測準確度，以及統整出影響匯率走勢之重要因素，也將模型的預測結果建構出投資組合，並使用評量指標分析模型的績效表現。本研究也利用實證上有效的因子，如：利差因子、動能因子、價值因子、市場因子，將此四個因子同時納入機器學習模型及羅吉斯回歸中，比較四因子使用不同方式做預測的差異，也將預測結果建構成投資組合，與實證上有超額報酬的外匯交易策略做比較。實證結果顯示，機器學習模型在預測匯率變動方向上比羅吉斯回歸佳，顯示機器學習確實有分類預測能力；使用六十個因子的預測準確度優於使用四個因子，顯示較多因子能涵蓋更多資訊，可得到更好預測結果；三種機器學習模型的預測準確度差異不大。在模型之重要因子方面，三種機器學習模型中，政府效能都是相當重要的因子，顯示一國政府治理、政治局勢是否穩定是影響匯率重要因素。在投資組合績效表現上，四因子納入機器學習模型優於實證上有效的單一策略，顯示多因子機器學習模型建構出來的策略比單一因子更好且更穩定；另外，也發現GBDT與XGBoost的投資組合有較高累積報酬，表現優於隨機森林。 zh_TW dc.description.abstract (摘要) Forex rates can be predicted by analyzing factors, like interest rate decisions, economic data, and political situations. If companies can have better opinions about forex rates, they can use less cost for hedging or speculators can earn returns from some strategies. In this study, I used factors in machine learning models to predict forex rates.In this study, Random Forest, Gradient Boosting Decision Tree and Extreme Gradient Boosting are used to predict forex rates. Forecast accuracy are compared in three models and important factors are found out. Predictions are formed a portfolio and the performance are compared between models. I also use the factors of Carry, Market, Momentum and Value in machine learning models and Logistic Regression to compare predictions and performance of trading strategies between models.Results show machine learning models are better than logistic regression in predicting the direction of forex changes. Besides, the accuracy of sixty factors is better than four factors. However, the accuracy between three machine learning models is not much different. In the feature importance, I found factors related to governance are important in all machine learning models. In the portfolio performance, using four factors is better than a single strategy, showing that a portfolio constructed by multi-factor machine learning model is better than a single factor. Besides, the study found GBDT and XGBoost are better than Random Forest in performance. en_US dc.description.tableofcontents 摘要 IIABSTRACT III目次 IV表次 V圖次 VI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 論文架構 3第二章 文獻回顧 4第一節 傳統模型文獻回顧 4第二節 機器學習模型文獻回顧 6第三章 樣本選擇及因子建構 9第一節 樣本選擇 9第二節 因子建構 12第四章 研究方法 20第一節 傳統模型 20第二節 機器學習模型 21第五章 實證結果分析 31第一節 評估模型預測表現-以AUC為標準 31第二節 模型之預測表現比較 32第三節 模型之因子重要性比較 35第四節 模型之交易策略績效比較 38第六章 結論與建議 42第一節 結論 42第二節 未來建議 43參考文獻 44 zh_TW dc.format.extent 2014191 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0108352023 en_US dc.subject (關鍵詞) 外匯交易策略 zh_TW dc.subject (關鍵詞) 機器學習 zh_TW dc.subject (關鍵詞) 隨機森林 zh_TW dc.subject (關鍵詞) 梯度提升樹 zh_TW dc.subject (關鍵詞) 極限梯度提升樹 zh_TW dc.subject (關鍵詞) Forex strategy en_US dc.subject (關鍵詞) Machine learning en_US dc.subject (關鍵詞) Random Forest en_US dc.subject (關鍵詞) Gradient Boosting Decision Tree en_US dc.subject (關鍵詞) Extreme Gradient Boosting en_US dc.title (題名) 應用機器學習於外匯超額報酬預測 zh_TW dc.title (題名) Applying Machine Learning for the Prediction of Foreign Exchange Return en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] 郭秀樺. (2018). 外匯報酬之利差、動能及價值交易策略成因分析. 國立政治大學碩士學位論文.[2] 林庭陞. (2020). 機器學習匯率定價投資組合. 國立政治大學碩士學位論文.[3] Acuña, E. & Rodriguez, C. (2004). The Treatment of Missing Values and its Effect on Classifier Accuracy. In: Banks D., McMorris F.R., Arabie P., Gaul W. (eds) Classification, Clustering, and Data Mining Applications. Studies in Classification, Data Analysis, and Knowledge Organisation. Springer, Berlin, Heidelberg.[4] Alsaleem, M.Y.A., & Hasoon, S.O. (2020). Comparison of dt& gbdt algorithms for predictive modeling of currency exchange rates. EUREKA: Physics and Engineering.[5] Asness, C.S., Moskowitz, T.J., & Pedersen, L.H. (2013). Value and Momentum Everywhere. The Journal of Finance, 68(3), 929-985.[6] Berg, K.A. & Mark, N.C. (2018). Measures of global uncertainty and carry-trade excess returns. Journal of International Money and Finance, 88, 212-227.[7] Booth, A., Gerding, E., & McGroarty, F. (2014). Automated trading with performance weighted random forests and seasonality. Expert Systems with Applications, 41(8), 3651-3661.[8] Bousbaa, Z., Chihab, M., Bencharef, O., & Zitiet, S. (2019). Algo-Trading Strategy for Intraweek Foreign Exchange Speculation Based on Random Forest and Probit Regression. Applied Computational Intelligence and Soft Computing.[9] Breiman, L. (2001). Random Forests. Machine Learning, 45, 5-32.[10] Brunnermeier, M.K., Nagel, S., & Pedersen, L.H. (2008). Carry Trades and Currency Crashes. NBER Macroeconomics Annual, 23.[11] Burnside, C., Eichenbaum, M., & Rebelo, S. (2011). Carry Trade and Momentum in Currency Markets. Annual Review of Financial Economics, 3, 511-535.[12] Chan, L.K.C., Jegadeesh, N., & Lakonishok, J. (1995). Evaluating the performance of value versus glamour stocks the impact of selection bias. Journal of Financial Economics, 38(3), 269-296.[13] Chen, C.N., & Lin, C.H. (2020). The sources of pricing factors underlying the cross-section of currency returns. The Quarterly Review of Economics and Finance, 77, 250-265.[14] Chen, T. & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. KDD `16: The 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco California, USA, 785-794.[15] Dey, S., Kumar, Y., Saha, S. & Basak, S. (2016). Forecasting to Classification: Predicting the direction of stock market price using Xtreme Gradient Boosting.[16] Du, W., Tepper, A., & Verdelhan, A. (2018). Deviations from covered interest rate parity. The Journal of Finance, 73(3), 915-957.[17] Fama, E.F., & French, K.R. (1992). The Cross-Section of Expected Stock Returns. The Journal of Finance, 47(2), 427-465.[18] Ferrari, M., Kearns, J. & Schrimpf, A. (2021). Monetary policy’s rising FX impact in the era of ultra-low rates. Journal of Banking & Finance, 129.[19] Filardo, A.J., Mohanty, M.S. & Moreno, R. (2012). Central Bank and Government Debt Management: Issues for Monetary Policy. BIS Paper No. 67d.[20] Friedman, J.H. (2001). Greedy Function Approximation: A Gradient Boosting Machine. The Annals of Statistics, 29(5), 1189-1232.[21] Galeshchuk, S. & Mukherjee, S. (2017). Deep networks for predicting direction of change in foreign exchange rates. Intelligent Systems in Accounting Finance & Management, 24(3).[22] Gu, S., Kelly, B. & Xiu, D. (2020). Empirical Asset Pricing via Machine Learning. The Review of Financial Studies, 33(5), 2223-2273.[23] Islam, S.F.N., Sholahuddin, A. & Abdullah, A.S. (2021). Extreme gradient boosting (XGBoost) method in making forecasting application and analysis of USD exchange rates against rupiah. Tenth International Conference and Workshop on High Dimensional Data Analysis (ICW-HDDA-X) 12-15 October 2020 in Sanur-Bali, Indonesia, 1722.[24] 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.[25] Kroencke, T.A., Schindler, F., & Schrimpf, A. (2014). International diversification benefits with foreign exchange investment styles. Review of Finance, 18(5), 1847- 1883.[26] Lintner, J. (1965). Security Prices, Risk, and Maximal Gains from Diversification. The Journal of Finance, 20(4), 587-615.[27] Liu, H. (2020). Stock Selection Strategy Based on Support Vector Machine and eXtreme Gradient Boosting Methods. 2020 the 4th International Conference on Big Data Research, 36-39.[28] Lukas, M., & Mark, P.T. (2007). The Obstinate Passion of Foreign Exchange Professionals: Technical Analysis. Journal of Economic. Literature, 45(4), 936-972.[29] Lustig, H., Roussanov, N., & Verdelhan, A. (2011). Common Risk Factor in Currency Markets. Review of Financial Studies. 24(11), 3731-3777.[30] Maragoudakis, M. & Serpanos, D. (2010). Towards Stock Market Data Mining Using Enriched Random Forests from Textual Resources and Technical Indicators.[31] Menkhoff, L., Sarno, L., Shmeling, M., & Schrimpf, A. (2012). Currency Momentum Strategies. Journal of Financial Economics, 106(3), 660-684.[32] Okunev, J., & White, D. (2003). Do Momentum-Based Strategies Still Work in Foreign Currency Markets? The Journal of Financial and Quantitative Analysis, 38(2), 425-447.[33] Qian, B. & Rasheed, K. (2010). Foreign exchange market prediction with multiple classifiers. Journal of Forecasting, 29(3), 271-284.[34] Qin, Q., Wang, Q.G., Li, J. & Ge, S.S. (2013). Linear and Nonlinear Trading Models with Gradient Boosted Random Forests and Application to Singapore Stock Market. Journal of Intelligent Learning Systems and Applications, 5(1).[35] Raza, A. (2015). Are Value Strategies Profitable in the Foreign Exchange Market? 28th Australasian Finance and Banking Conference[36] Schut, F.G.B., van Rijn, J.N., & Hoos, H.H. (2015). Towards Automated Technical Analysis for Foreign Exchange Data.[37] Sharpe, W.F. (1964). Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk. The Journal of Finance, 19(3), 425-442. zh_TW dc.identifier.doi (DOI) 10.6814/NCCU202100724 en_US
