本文共有三個實驗命題，命題一，研究深度強化學習模型在不同市場情境下如何應變。命題二，比較深度強化學習資產配置模型與加入市場分數之資產配置模型之風險報酬。命題三，探討加入市場分數之資產配置模型與加入無風險資產之資產配置模型差異。以三命題探索不同市場趨勢與不同資產池對於深度強化學習資產配置模型之各類比較，命題一研究顯示深度強化學習資產配置模型在市場趨勢屬於恐慌時期與熊市時皆能有效改善風險報酬率；命題二研究成果顯示將資產配置模型拆成資產分數與市場分數兩部分，能在有效降低風險同時保有一定的獲利能力，風險報酬率更勝單純資產配置模型。命題三研究成果顯示加入市場分數與於資產池中加入無風險資產之資產模型各有優缺點，然長期來看加入市場分數較能在承受相同風險條件下追求更優的獲利率。三命題皆以風險與報酬指標來比較不同資產配置模型優劣，期望能建構出穩定獲利資產配置模型。

There are three experimental purposes in this paper. First, how will the DRL asset management model respond to different market trends？ Second, compare the risk-reward of the DRL asset management model with the DRL asset management model added market score. Third, discuss the difference between the asset management model adding market scores and the asset management model adding risk-free assets. Explore various comparisons of different market trends and different asset pools for the DRL asset management model with three propositions. The results revealed that the DRL model can effectively improve the risk-reward ratio during the great depression and bear market. The research results of proposition 2 show that splitting the asset management model into two parts, the asset score and the market score, can effectively reduce risks while maintaining certain profitability, and the risk-reward ratio is better than the traditional asset management model. The research results of proposition 3 show that adding market scores and adding risk-free assets to the asset pool have its own advantages and disadvantages. However, in the long term, adding market scores can pursue better profitability under the same risk conditions.

[1] Chen, C., Zhao, L., Bian, J., Xing, C., & Liu, T. Y. (2019). Investment behaviors can tell what inside: Exploring stock intrinsic properties for stock trend prediction. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 2376-2384)
[2] Darapaneni, N., Basu, A., Savla, S., Gururajan, R., Saquib, N., Singhavi, S., ... & Paduri, A. R. (2020). Automated portfolio rebalancing using Q-learning. In 2020 11th IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) (pp. 0596-0602).
[3] Deng, Y., Bao, F., Kong, Y., Ren, Z., & Dai, Q. (2016). Deep direct reinforcement learning for financial signal representation and trading. IEEE transactions on neural networks and learning systems, 28(3) (pp.653-664).
[4] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8) (pp. 1735-1780).
[5] Hoseinzade, E., & Haratizadeh, S. (2019). CNNpred: CNN-based stock market prediction using a diverse set of variables. Expert Systems with Applications, 129 (pp. 273-285).
[6] Jiang, Z., Xu, D., & Liang, J. (2017). A deep reinforcement learning framework for the financial portfolio management problem. arXiv preprint arXiv:1706.10059.
[7] Jin, O., & El-Saawy, H. (2016). Portfolio management using reinforcement learning. Stanford University.
[8] Lintner, J. (1975). The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. In Stochastic optimization models in finance (pp. 131-155).
[9] Liu, F., Li, Y., Li, B., Li, J., & Xie, H. (2021). Bitcoin transaction strategy construction based on deep reinforcement learning. Applied Soft Computing, 113, 107952.
[10] Liu, X. Y., Yang, H., Chen, Q., Zhang, R., Yang, L., Xiao, B., & Wang, C. D. (2020). FinRL: A deep reinforcement learning library for automated stock trading in quantitative finance. arXiv preprint arXiv:2011.09607.
[11] Nelson, D. M., Pereira, A. C., & De Oliveira, R. A. (2017). Stock market's price movement prediction with LSTM neural networks. In 2017 International joint conference on neural networks (IJCNN) (pp. 1419-1426).
[12] Ozbayoglu, A. M., Gudelek, M. U., & Sezer, O. B. (2020). Deep learning for financial applications: A survey. Applied Soft Computing, 93, 106384.
[13] Roondiwala, M., Patel, H., & Varma, S. (2017). Predicting stock prices using LSTM. International Journal of Science and Research (IJSR), 6(4) (pp. 1754-1756).
[14] Solorio-Fernández, S., Carrasco-Ochoa, J. A., & Martínez-Trinidad, J. F. (2020). A review of unsupervised feature selection methods. Artificial Intelligence Review, 53(2) (pp. 907-948).
[15] Statman, M. (1987). How many stocks make a diversified portfolio?. Journal of financial and quantitative analysis, 22(3) (pp. 353-363).
[16] Understanding LSTM Networks. http://colah.github.io/posts/2015-08-Understanding-LSTMs. [Online; accessed 9-NOV-2022]
[17] Vargas, M. R., De Lima, B. S., & Evsukoff, A. G. (2017). Deep learning for stock market prediction from financial news articles. In 2017 IEEE international conference on computational intelligence and virtual environments for measurement systems and applications (CIVEMSA) (pp. 60-65)
[18]Wang, B., & Zhang, X. Deep Learning Applying on Stock Trading. Stanford University.
[19] Wang, J., Zhang, Y., Tang, K., Wu, J., & Xiong, Z. (2019). Alphastock: A buying-winners-and-selling-losers investment strategy using interpretable deep reinforcement attention networks. In Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining (pp. 1900-1908).
[20] Wang, Z., Huang, B., Tu, S., Zhang, K., & Xu, L. (2021). DeepTrader: a deep reinforcement learning approach for risk-return balanced portfolio management with market conditions Embedding. In Proceedings of the AAAI Conference on Artificial Intelligence, 35(1) (pp. 643-650).
[21] What are neural networks? https://www.ibm.com/cloud/learn/neural-networks. [Online; accessed 9-NOV-2022]
[22] Wu, X., Chen, H., Wang, J., Troiano, L., Loia, V., & Fujita, H. (2020). Adaptive stock trading strategies with deep reinforcement learning methods. Information Sciences, 538 (pp. 142-158).