本論文是建構在Fama French 因子模型上，利用深度學習的優勢，找出公司特徵值之間的非線性關係。目的是透過機器學習，降低公司橫斷面預測報酬與實際報酬的誤差值。

在深度學習模型裡，我們利用個股特徵值進行排序，依市值加權的方式組合成一個買進-放空策略，形成一個可交易的因子。除了深度學習外，我們還另外提供了PCA的方法來組合因子，來當作對照組。

在這次的研究過程裡，個股資料主要來自於台灣上市股票報酬以及公司財報當作特徵值，針對2000年到2018年的台灣股票市場來進行相關回測。深度學習建構出的因子確實有解釋能力，而且無論是在解釋能力還是捕捉異常報酬的個數上，都比PCA在樣本外的表現來得穩定。

最後，去探討因子之間吸收效果的存在，和PCA在樣本外的結果不穩定問題，並給出許多未來仍可以改善的方針與建議。

This study is constructed based on the Fama and French factor model, using the advantages of deep learning to determine the nonlinear relationship between firm characteristics. The purpose of the study is to minimize the pricing error between cross-sectional forecasted return and actual return.

Deep learning provides a framework for characteristics-based factor modeling in empirical asset pricing. We use security sorting based on firm characteristics, combining them into long-short factors weighted by market capitalization to create a tradable factor. Besides, we also provide an additional PCA method to form factors as a comparison group.

For cross-sectional return prediction, we study monthly Taiwan equity returns and firm characteristics from 2000 to 2018 with a universe of 1000 stocks. The factors constructed through deep learning have actual explanatory power of return, and the deep learning model is considerably stable in the out-sample period, both in explanatory power and ability to grab anomalies.

Finally, we further explore the absorb effect between the factors and discover that PCA results have some defects in the out-sample period. Besides, we also give some guidelines and suggestions that could improve the model in the future.

論文審定書 i
摘要 ii
ABSTRACT iii
I. INTRODUCTION 1
1.1 Background Information 1
1.2 Purpose of Research 3
1.3 Research Framework 4
II. LITERATURE REVIEW 5
2.1 Modern Portfolio Theory 5
2.2 Principal Component Analysis 6
2.3 Deep Learning 7
III. METHODOLOGY 10
3.1 Deep Characteristics 11
3.2 Deep Factors 13
3.3 Minimizing Mispricing Alphas 17
3.4 Comparison Method: Principal Component Analysis 19
IV. EMPIRICAL RESULTS 20
4.1 Data 20
4.2 Train-Validation-Test Design 21
4.3 Out-of-sample Forecasts 23
4.4 Dissecting Anomalies 26
4.5 Absorbing Effect 28
V. CONCLUSION 32
5.1 Conclusion 32
5.2 Suggestions for Future Research 34
VI. REFERENCES 36
Appendix A: Characteristics and Anomaly Summary 40
Appendix B: Optimization Details 42
A. Activation 42
B. Optimizer function 43
C. Parameter 45
Appendix C: Principal Component Analysis 50

1. Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of Finance, 52(1), 57-82.
2. Chamberlain, G., & Rothschild, M. (1983). Arbitrage, Factor Structure, and Mean& Variance Analysis on Large Asset Marketsn. Econometrica, 51, 1281.
3. Connor, G. (1995). The three types of factor models: A comparison of their explanatory power. Financial Analysts Journal, 51(3), 42-46.
4. Connor, G., & Korajczyk, R. A. (1986). Performance measurement with the arbitrage pricing theory: A new framework for analysis. Journal of Financial Economics, 15(3), 373-394.
5. Connor, G., & Korajczyk, R. A. (1988). Risk and return in an equilibrium APT: Application of a new test methodology. Journal of Financial Economics, 21(2), 255-289.
6. Cox, J. C., & Ross, S. A. (1976). A survey of some new results in financial option pricing theory. The Journal of Finance, 31(2), 383-402.
7. Duchi, J., Hazan, E., & Singer, Y. (2011). Adaptive subgradient methods for online learning and stochastic optimization. Journal of Machine Learning Research, 12(Jul), 2121-2159.
8. Fama, E. F., & MacBeth, J. D. (1973). Risk, return, and equilibrium: Empirical tests. Journal of Political Economy, 81(3), 607-636.
9. Fama, E. F., & French, K. R. (1992). The cross‐section of expected stock returns. The Journal of Finance, 47(2), 427-465.
10. Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33(1), 3-56.
11. Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33(1), 3-56.
12. Fama, E. F., & French, K. R. (1997). Industry costs of equity. Journal of Financial Economics, 43(2), 153-193.
13. Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1-22.
14. Fama, E. F., & French, K. R. (2018). Choosing factors. Journal of Financial Economics, 128(2), 234-252.
15. Feng, G., Giglio, S., & Xiu, D. (2017). Taming the factor zoo. Fama-Miller Working Paper, 24070.
16. Feng, G., Polson, N. G., & Xu, J. (2019). Deep learning in asset pricing. Technical report, City University of Hong Kong.
17. Green, J., Hand, J. R., & Zhang, X. F. (2017). The characteristics that provide independent information about average us monthly stock returns. The Review of Financial Studies, 30(12), 4389-4436.
18. Freyberger, J., Neuhierl, A., & Weber, M. (2017). Dissecting characteristics nonparametrically (No. w23227). National Bureau of Economic Research.
19. Heaton, J. B., Polson, N. G., & Witte, J. H. (2017). Deep learning for finance: deep portfolios. Applied Stochastic Models in Business and Industry, 33(1), 3-12.
20. Harvey, C. R., Liu, Y., & Zhu, H. (2016). … and the cross-section of expected returns. The Review of Financial Studies, 29(1), 5-68.
21. Heaton, J., Polson, N. G., & Witte, J. H. (2016). Deep Learning in Finance. arXiv preprint arXiv:1602.06561.
22. Kelly, B. T., Pruitt, S., & Su, Y. (2019). Characteristics Are Covariances: A unified model of risk and return. Journal of Financial Economics.
23. Kingma, D., & Ba, J. (2015). Adam: a method for stochastic optimization (2014). arXiv preprint arXiv:1412.6980, 15.
24. LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep Learning. Nature, 521(7553), 436.
25. Lettau, M., & Pelger, M. (2018). Factors that fit the time series and cross-section of stock returns (No. w24858). National Bureau of Economic Research.
26. Lintner, J. (1965). Security prices, risk, and maximal gains from diversification. The Journal of Finance, 20(4), 587-615.
27. Markowitz, H. (1952). Portfolio selection. The Journal of Finance, 7(1), 77-91.
28. Merton, R. C. (1973). An intertemporal capital asset pricing model. Econometrica, 41(5), 867-887.
29. Merton, R. C. (1973). Theory of rational option pricing. Theory of Valuation, 229-288.
30. Polson, N. G., & Sokolov, V. (2017). Deep learning: a Bayesian perspective. Bayesian Analysis, 12(4), 1275-1304.
31. Sharpe, W. F. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk. The Journal of Finance, 19(3), 425-442.
32. Treynor, J. L. (1962). Jack Treynor's' Toward a Theory of Market Value of Risky Assets'. Available at SSRN 628187.
33. Tieleman, T., & Hinton, G. (2012). Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude. COURSERA: Neural networks for machine learning, 4(2), 26-31