有鑑於2020年新冠肺炎疫情這樣的黑天鵝事件，導致各國的股票市場嚴重下跌，即便是保守投資人也可能承受巨大的虧損，因此本研究希望可以針對保守投資人建構穩定低波動且獲利高的投資組合。本研究使用Blackrock, Vanguard, State Street Global Advisors美國前三大基金管理公司所發行的ETF作為研究樣本，計算2011年1月至2015年12月共5年的績效，並依據不同類型的ETF，給予不同的篩選績效條件，接著為篩選出的ETF建立特徵變數，特徵變數分成三大類，分別為技術指標、流動性指標、總體經濟指標，將建好的特徵變數輸入XGBoost模型做訓練，研究回測期間為2016年1月至2020年12月，使用移動窗格的訓練方式，訓練期資料為5年，預測一個月後的月波動度，每個月重新訓練一次模型且本研究每檔ETF為分開做訓練。研究結果顯示等風險權重投資組合的夏普比率及最大回撤率皆表現得比等權重投資組合及大盤指數好，而在股票市場遭遇系統性風險時，更可以觀察到等風險權重投資組合的優勢。另外，使用預測一個月後的月波動度所組成的等風險權重投資組合會比使用前一個月的月波動度所組成的等風險權重投資組合績效表現較好。

In view of the black swan incident such as the COVID-19 pandemic in 2020, which has caused severe declines in the stock markets of various countries, even conservative investors may suffer huge losses. Therefore, this study hopes to build a portfolio which is low volatility and high profits for conservative investors. This study uses ETFs issued by the top three fund management companies in the United States of Blackrock, Vanguard, and State Street Global Advisors as the research sample. It calculates the performance of the five-year period from January 2011 to December 2015. According to the performance of ETFs, the eligible ETFs are selected, and then variables are established for the selected ETFs. Variables are divided into three categories, technical indicators, liquidity indicators, and economic indicators. Input the established variables into the XGBoost model for training. The backtesting period is from 2016 to 2020. The training period data is 5 years. The monthly volatility after one month is predicted. The model is retrained every month and each ETF is trained separately.The results of the research show that the sharpe ratio and maximum drawdown of equal-risk-weighted portfolios are better than other portfolios. In addition, an equal-risk-weighted portfolio composed of monthly volatility predicted one month later will perform better than an equal-risk-weighted portfolio composed of monthly volatility in the previous month.

論文審定書 i
摘要 ii
Abstract iii
目錄 iv
圖次 vi
表次 vii
第一章、 緒論 1
第一節、 研究動機 1
第二節、研究目的 2
第三節、研究架構 2
第二章、 文獻回顧 4
第一節、 投資組合 4
第二節、 總經指標與技術指標在預測股票市場之應用 5
第三節、 機器學習模型預測股票市場應用 6
第四節、小結 7
第三章、 研究方法 8
第一節、 研究流程 8
第二節、 研究資料 10
第三節、 篩選ETF 12
第四節、 特徵變數 15
第五節、 月波動度計算公式 24
第六節、 機器學習建構投組 24
第四章、 實證結果 28
第一節、 預測誤差 28
第二節、 特徵重要度 30
第三節、 投組績效 32
第五章、 研究結論 37
第一節、結論 37
第二節、研究建議 38
參考文獻 39

Ballings, M., Van den Poel, D., Hespeels, N., & Gryp, R. (2015). Evaluating multiple classifiers for stock price direction prediction. Expert systems with Applications, 42(20), 7046-7056.
Basak, S., Kar, S., Saha, S., Khaidem, L., & Dey, S. R. (2019). Predicting the direction of stock market prices using tree-based classifiers. The North American Journal of Economics and Finance, 47, 552-567.
Chaves, D., Hsu, J., Li, F., & Shakernia, O. (2011). Risk parity portfolio vs. other asset allocation heuristic portfolios. The Journal of Investing, 20(1), 108-118.
Chen, S. S. (2009). Predicting the bear stock market: Macroeconomic variables as leading indicators. Journal of Banking & Finance, 33(2), 211-223.
Chen, T., & Guestrin, C. (2016, August). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining (pp. 785-794).
Dai, Z., Zhou, H., Dong, X., & Kang, J. (2020). Forecasting Stock Market Volatility: A Combination Approach. Discrete Dynamics in Nature and Society, 2020.
Fadlalla, A., & Amani, F. (2014). Predicting next trading day closing price of Qatar exchange index using technical indicators and artificial neural networks. Intelligent Systems in Accounting, Finance and Management, 21(4), 209-223.
Goetzmann, W. N., & Kumar, A. (2008). Equity portfolio diversification. Review of Finance, 12(3), 433-463.
Kong, A., Zhu, H., & Azencott, R. (2020). Predicting intraday jumps in stock prices using liquidity measures and technical indicators. Journal of Forecasting.
Liew, J. K. S., & Mayster, B. (2017). Forecasting etfs with machine learning algorithms. The Journal of Alternative Investments, 20(3), 58-78.
Liu, L., & Pan, Z. (2020). Forecasting stock market volatility: The role of technical variables. Economic Modelling, 84, 55-65.
Maillard, S., Roncalli, T., & Teiletche, J. (2008). On the properties of equally-weighted risk contributions portfolios. SSRN.
Markowitz, H. (1952). Portfolio selection. The Journal of Finance, 7, 77-91.
Nevasalmi, L. (2020). Forecasting multinomial stock returns using machine learning methods. The Journal of Finance and Data Science, 6, 86-106.
Shynkevich, Y., McGinnity, T. M., Coleman, S. A., Belatreche, A., & Li, Y. (2017). Forecasting price movements using technical indicators: Investigating the impact of varying input window length. Neurocomputing, 264, 71-88.
Simon, D. P., & Wiggins III, R. A. (2001). S&P futures returns and contrary sentiment indicators. Journal of Futures Markets: Futures, Options, and Other Derivative Products, 21(5), 447-462.
Zhang, X., & Chen, W. (2019, July). Stock Selection Based on Extreme Gradient Boosting. In 2019 Chinese Control Conference (CCC) (pp. 8926-8931). IEEE.