本研究主要是透過Ranking SVM模型來預測股票的排名順序，從而找出排名靠前的股票投資並回測。方法則是利用機器學習中「長短期記憶模型－LSTM」和「自編碼器－AutoEncoder」，將代表股價走勢的技術指標資料做同化，並且轉換成高維度的特徵向量。接著透過「支持向量機-SVM」將特徵向量做兩兩股票漲跌幅的預測分類，並選出預測最佳的股票做投資並回測。由於股票的預測排名與實際排名存在落差，我們透過「平均－變異數優化法(MVO)」，找出一段時間內、一群股票中，其預測排名差負向變化率平均值最大的股票，並以此股票投資並看回報率。為了優化單一MVO周期模型的回報率，我們組合不同周期的MVO模型，並以此模型來推薦股票並投資。最終，我們使用組合型周期的MVO模型得到的累積回報率比元大ETF50的累積回報率要來的更好。

The purpose of this study was to predict the ranking of stocks by using the Ranking SVM model and got a good accumulation return. Long Short-Term Memory (LSTM)-based AutoEncoder model was applied for data assimilation and higher-dimensional feature projecting. The training data was arranged by the pairwise method and was input to the SVM model for the classification of return comparison from every two stocks. The top-ranking stocks from prediction were used for investment; On the other hand, because of the ranking-prediction error from the classifier, Mean-Variance Optimization(MVO) was applied for post-processing. By choosing the minimum variance of the ranking-prediction error, the
recommended investment stock could be found in each of the MVO models with different periods. In advance, each of the MVO models with different periods was chosen and combined into a composted MVO model for a better return. In the end, the accumulation return from composted MVO model was superior to the ETF50 ones.

[1] Z. Bodie and A. Kane. Investments. McGraw Hill, 2020.
[2] H. M. Markowitz. Portfolio selection. Journal of Finance, 7:77–91, 1952.
[3] Y. Kroll, H. Levy, and A. Rapoport. Experimental tests of the separation theorem
and the capital asset pricing model. The American Economic Review, pages 500–519,
1988.
[4] E. F. Fama and K. R. French. The capital asset pricing model: Theory and evidence.
Journal of Economic Perspectives, 18:25–46, 2004.
[5] E. F. Fama and K. R. French. Common Risk Factors in the Returns on Stocks and
Bonds. University of Chicago Press, 2021.
[6] E. F. Fama and K. R French. The Cross-Section of Expected Stock Returns. University
of Chicago Press, 2021.
[7] J Patel, S Shah, P Thakkar, and K Kotecha. Predicting stock and stock price
index movement using trend deterministic data preparation and machine learning
techniques. Expert Systems with Applications, 42:259–268, 2015.
[8] V.-D. Ta, C.-M. Liu, and D. Addis. Prediction and portfolio optimization in quanti-
tative trading using machine learning techniques. In Proceedings of the Ninth Inter-
national Symposium on Information and Communication Technology, pages 98–105,
2018.
[9] T. T.-L. Chong and W.-K. Ng. Technical analysis and the London stock exchange:
testing the MACD and RSI rules using the ft30. Applied Economics Letters, 15:1111–
1114, 2008.
[10] R. Rosillo, D. De la Fuente, and J. Brugos. Technical analysis and the spanish stock
exchange: testing the rsi, MACD, momentum and stochastic rules using spanish
market companies. Applied Economics, 45:1541–1550, 2013.
[11] Y. Ma, R. Han, and W. Wang. Prediction-based portfolio optimization models using
deep neural networks. IEEE Access, 8:115393–115405, 2020.
38
[12] S. Obeidat, D. Shapiro, M. Lemay, M. K. MacPherson, and M. Bolic. Adaptive
portfolio asset allocation optimization with deep learning. International Journal on
Advances in Intelligent Systems, 11:25–34, 2018.
[13] C.-M. Lin, J.-J. Huang, M. Gen, and G.-H. Tzeng. Recurrent neural network for
dynamic portfolio selection. Applied Mathematics and Computation, 175:1139–1146,
2006.
[14] X. Y. Fu, J. H. Du, Y. F. Guo, M. W. Liu, T. Dong, and X. W. Duan. A machine
learning framework for stock selection. arXiv preprint arXiv:1806.01743, 2018.
[15] V. Kedia, Z. Khalid, S. Goswami, N. Sharma, and K. Suryawanshi. Portfolio gener-
ation for indian stock markets using unsupervised machine learning. In 2018 Fourth
ICCUBEA, pages 1–5. IEEE, 2018.
[16] Z. Zhang, S. Zohren, and S. Roberts. Deep learning for portfolio optimization. The
Journal of Financial Data Science, 2:8–20, 2020.
[17] C. Krauss, X. A. Do, and N. Huck. Deep neural networks, gradient-boosted trees,
random forests: Statistical arbitrage on the S&P 500. European Journal Operation
Research, 259:689–702, 2017.
[18] E. Tas and Ayca H. Atli. Stock price ranking by learning pairwise preferences.
Computational Economics, pages 1–16, 2022.
[19] T. Engin. A single pairwise model for classification using online learning with kernels.
Hacettepe Journal of Mathematics and Statistics, 46:547–557, 2017.
[20] M. Chang. Application of learning to rank and autoencoder hybrid technology in
portfolio strategy. Master’s thesis, National Central University, 2021.
[21] O. Jin and H. El-Saawy. Portfolio management using reinforcement learning.
Preprint, 2016.
[22] Z. Jiang, D. Xu, and J. Liang. A deep reinforcement learning framework for the
financial portfolio management problem. arXiv preprint arXiv:1706.10059, 2017.
[23] Z. Jiang and J. Liang. Cryptocurrency portfolio management with deep reinforcement
learning. In 2017 Intelligent Systems Conference (IntelliSys), pages 905–913. IEEE,
2017.
39
[24] T. Singh, R. Kalra, S. Mishra, M. Kumar, et al. An efficient real-time stock prediction
exploiting incremental learning and deep learning. Evolving Systems, pages 1–19,
2022.
[25] S. Rathor and S. Agrawal. A robust model for domain recognition of acoustic com-
munication using bidirectional LSTM and deep neural network. Neural Computing
and Applications, 33:11223–11232, 2021.
[26] X. Du, H. Zhang, H. Van Nguyen, and Z. Han. Stacked LSTM deep learning model for
traffic prediction in vehicle-to-vehicle communication. In 2017 IEEE 86th Vehicular
Technology Conference (VTC-Fall), pages 1–5. IEEE, 2017.
[27] Y. Wu, M. Yuan, S. Dong, L. Lin, and Y. Liu. Remaining useful life estimation of
engineered systems using vanilla LSTM neural networks. Neurocomputing, 275:167–
179, 2018.
[28] A. Graves and J. Schmidhuber. Framewise phoneme classification with bidirectional
LSTM and other neural network architectures. Neural Networks, 18:602–610, 2005.
[29] B. Li, Q. Wang, and J. Hu. Feature subset selection: a correlation-based SVM filter
approach. IEEJ Transactions on Electrical and Electronic Engineering, 6:173–179,
2011.
[30] B. J. Frey and D. Dueck. Clustering by passing messages between data points.
science, 315:972–976, 2007.
[31] H. Xing, M. Ha, B. Hu, and D. Tian. Linear feature-weighted support vector machine.
Fuzzy Information and Engineering, 1:289–305, 2009.
[32] B. Chen, F. Sun, and J. Hu. Local linear multi-SVM method for gene function
classification. In 2010 Second World Congress on Nature and Biologically Inspired
Computing (NaBIC), pages 183–188. IEEE, 2010.
[33] A. Fan and M. Palaniswami. Stock selection using support vector machines. In
IJCNN’01. International Joint Conference on Neural Networks. Proceedings (Cat.
No. 01CH37222), volume 3, pages 1793–1798. IEEE, 2001.
[34] V. Vapnik. The Nature of Statistical Learning Theory. Springer science & business
media, 1999.
40
[35] K. Veropoulos, C. Campbell, N. Cristianini, et al. Controlling the sensitivity of
support vector machines. In Proceedings of the international joint conference on AI,
volume 55, page 60. Stockholm, 1999.
[36] T. Tantisripreecha and N. Soonthomphisaj. Stock market movement prediction using
lda-online learning model. In 2018 19th IEEE/ACIS International Conference on
Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed
Computing (SNPD), pages 135–139. IEEE, 2018.