本研究主要是以機器學習模型為基礎，透過機器學習方法找出是否有潛在的因子能夠讓傳統上使用的因子外對外匯的超額報酬提供更高的解釋力，本研究使用樹模型、梯度提升、具隱藏層的神經網路以及隨機森林模型。本研究先在樣本期間(1997/01 至 2019/05)以HML投組法將19國匯率資料建構出利差、動能以及價值交易策略因子來取得傳統模型使用的因子。除了傳統上常使用的因子外，我們同時加入了其他的總經因子以及個別國家因子進入我們的模型之中，透過因子重要性分析我們發現不同的因子對於不同的模型會有不一樣的影響程度，但是除了市場因子外，並未有一個新加入的因子能夠顯著影響到所有的機器學習模型，故我們在結論處提出未來能夠改進的方向。

This paper tries to find some latent factors that can help us explain excess foreign return efficiently except using traditional factors which are market factor, value factor, momentum factor, and carry trade factor. We use three kinds of machine learning models in this paper which are random forest model, gradient boosting tree model, and neural network with hidden layer from one to five models. First, we use 19 countries’ foreign exchange data from Jan. 1997 to May. 2019 to build traditional factors by HML method. Then, we also put some macro factors and country-specific factors into machine learning models. Last, we specify which factor can affect the explanation ability separately by ranking variable importance in each model.

[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。國立政治大學金融研究所碩士論文，台北市。
[2] Barroso, P., & Santa-Clara, P. (2015). Beyond the carry trade: Optimal currency portfolios. The Journal of Financial and Quantitative Analysis, 50, 1037–1056.
[3] Brunnermeier, M. K., Nagel, S., & Pedersen, L. H. (2009). Carry trades and currency crashes. NBER Macroeconomics Annual, 23, 313–348.
[4] Butaru, F., Qingqing C., Brian C., Sanmay D., Andrew L., & Akhtar S., (2016), Risk and risk management in the credit card industry, Journal of Banking & Finance 72, 218~239.
[5] Chordia, T., & Shivakumar, L. (2002). Momentum, business cycle, and time-varying expected returns. The Journal of Finance, 62, 985–1019.
[6] Fama, E., & French, K. (1988). Business cycles and the behavior of metals prices. The Journal of Finance, 43(5), 1075–1093.
[7] Fama, E. F., & MacBeth, J. (1973). Risk, return and equilibrium: Empirical tests. The Journal of Political Economy, 81, 607–636.
[8] Freyberger, Joachim, Andreas N., and Michael W., (2017), Dissecting characteristics nonparametrically,Technical report, University of Wisconsin-Madison.
[9] Harvey, R., & Wayne F., (1999), Conditioning variables and the cross-section of stock returns, Journal of Finance 54, 1325~1360.
[10] Heaton, J., NG P., & JH W., (2016), Deep learning in finance
[11] Hutchinson, M., Andrew L., and Tomaso P., (1994), A nonparametric approach to pricing and hedging derivative securities via learning networks, The Journal of Finance 49, 851~889.
[12] Jingtao,Y. Yili L., & Chew T., (2000), Option price forecasting using neural networks, Omega 28, 455~466.
[13] Khandani, E., Adlar K., & Andrew L., (2010), Consumer credit-risk models via machine learning algorithms, Journal of Banking & Finance 34, 2767~2787.
[14] Lewellen, J. (2015), The cross-section of expected stock returns, Critical Finance Review 4, 1-44
[15] Lustig, H., Roussanov, N., & Verdelhan, A. (2011). Common risk factors in currency markets. The Review of Financial Studies, 24, 3731–3777.
[16] Menkhoff, L., Sarno, L., Schmeling, M., & Schrimpf, A. (2012b). Currency momentum strategies. Journal of Financial Economics, 106,660–684.
[17] Moritz, Benjamin, & Tom Z., (2016), Tree-based conditional portfolio sorts: The relation between past and future stock returns, Available at SSRN 2740751 .
[18] Rapach, E., Jack S., & Guofu Z., (2013), International stock return predictability: what is the role of the united states? The Journal of Finance 68, 1633~1662.
[19] Raza, A., Marshall, B. R., & Visaltanachoti, N. (2014). Is there momentum or reversal in weekly currency returns? Journal of International Money and Finance, 45,38–60.
[20] Shihao, G.,Bryan,K., & Dacheng, X.(2019). Empirical Asset Pricing via Machine Learning. NBER
[21] Sirignano, J., Apaar S., & Kay G., (2016), Deep learning for mortgage risk, Available at SSRN 2799443 .