台股自2020年開放零股交易以來，散戶交割戶數不斷創新高，「存股」演變成為全民運動。同時，機器學習在證券市場的應用雖可見於國外文獻，惟國內相關的實證研究中，對於價值投資的部分著墨較少。為了讓散戶投資人免於在新冠肺炎 (COVID-19) 疫情以及後疫情時代面臨低利率及高通膨導致實質薪資倒退的窘境，本研究即以價值投資選股哲學為據，透過當季台股基本面資料建立隨機森林模型，對於下一季的股價進行預測；同時藉由重要變數特徵作為價值選股依據篩選出符合條件的類股，以Markowitz(1952)預期報酬-變異數模型進行資產配置，並檢驗其報酬是否能優於同期大盤。實證結果發現，選擇「大型而卓越的企業」、「連續五年均獲利」、「連續五年均配息」且「價格合理」的公司所建構的投資組合在過去2013年至2019年期間的績效表現顯著優於台灣加權指數平均報酬。本研究以統計至2022年1月為止台灣股票市場市值前300大公司於2013Q1~2021Q4期間的基本面資料，排除金融類股與營建類股，以及剔除上市期間未滿5年的資料後，實際樣本包含226間公司，25個自變數，共8,136筆資料。

The number of retail settlement accounts has been skyrocketing since Taiwan authorities concerned opened up retail trading in 2020, causing “stock deposit” to evolve into a national movement; meanwhile, machine learning applied in the securities market can be seen in foreign literature, but value investment is rarely discussed in the part of relevant domestic empirical research. To help Taiwanese retail investors refraining from inflation and shrinking real wages in the COVID-19 pandemic and the post-pandemic era, this article is aiming to establish a random forest model based on the fundamental data of Taiwan corps in the current quarter, and predict the stock price of the next quarter. At the same time, the model utilizes the characteristics of important variables as the basis for value investing stock selection to formulate Markowitz (1952) optimized-portfolio, testing whether its return can outperform the broader market in the same period. The empirical results found that the portfolio constructed by companies that selected "large and outstanding companies", "profitable for five consecutive years", "distributed dividends for five consecutive years" and "reasonably priced" in the past 2013 to 2019 performed significantly better than the Taiwan Weighted Index Average Return. This research uses the fundamental data of the top 300 companies by market capitalization in Taiwan's stock market from 2013Q1 to 2021Q4, excluding financial stocks and construction stocks, and after excluding the listing period of less than 5 years, the actual sample includes 226 companies, 25 independent variables, a total of 8,136 records.

誌謝...............................................................II
摘要...............................................................III
Abstract...........................................................IV
第一章 緒論......................................................1
第二章 文獻探討..................................................3
第一節 價值投資相關文獻研究.......................................3
第二節 機器學習演算法相關文獻研究..................................5
第三節 研究假說..................................................7
第三章 研究方法..................................................8
第一節 研究架構..................................................8
第二節 選股規則說明..............................................10
第三節 樣本與資料來源............................................12
第四節 個案分析方法..............................................13
第四章 研究結果.................................................18
第一節 敘述性統計...............................................18
第二節 隨機森林訓練結果..........................................20
第三節 模型預測效果..............................................21
第四節 選股名單的決定............................................21
第五節 Markowitz 最適投資組合....................................23
第六節 績效回測..................................................24
第五章 結論與建議................................................26
第一節 研究結論..................................................26
第二節 研究限制與建議............................................27
參考文獻...........................................................28

1.Abarbanell, J. S., & Bushee, B. J. (1998). Abnormal returns to a fundamental analysis strategy. Accounting Review, 19-45.
2.Alberg, J., & Lipton, Z. C. (2017). Improving factor-based quantitative investing by forecasting company fundamentals. arXiv preprint arXiv:1711.04837.
3.Ball, R., & Brown, P. (1968). An empirical evaluation of accounting income numbers. Journal of accounting research, 159-178.
4.Beaver, W. H. (1968). The information content of annual earnings announcements. Journal of accounting research, 67-92.
5.Beaver, W. H., Clarke, R., & Wright, W. F. (1979). The association between unsystematic security returns and the magnitude of earnings forecast errors. Journal of accounting research, 316-340.
6.Boonpeng, S., & Jeatrakul, P. (2016, February). Decision support system for investing in stock market by using OAA-neural network. In 2016 Eighth International Conference on Advanced Computational Intelligence (ICACI) (pp. 1-6). IEEE.
7.Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.
8.Capaul, C., Rowley, I., & Sharpe, W. F. (1993). International value and growth stock returns. Financial Analysts Journal, 49(1), 27-36.
9.Arshanapalli, B., Daniel Coggin, T., & Doukas, J. (1998). Multifactor asset pricing analysis of international value investment strategies. Journal of Portfolio Management, 24(4), 10-+.
10.Dai, Y., & Zhang, Y. (2013). Machine learning in stock price trend forecasting. Stanford University, http://cs229. stanford. edu/proj2013/DaiZhang-MachineLearningInStockPriceTrendForecasting. pdf. Erişim Tarihi, 21, 2021.
11.Fama, E. F., & French, K. R. (1992). The cross‐section of expected stock returns. the Journal of Finance, 47(2), 427-465.
12.GOSSET, W. S. (2016). Gosset, william sealy. Encyclopedia of Mathematics.
13.Graham, B., Dodd, D. L. F., & Cottle, S. (1934). Security analysis (Vol. 452). New York: McGraw-Hill.
14.Graham, B. (1965). The intelligent investor. Prabhat Prakashan.
15.Giacomel, F., Galante, R., & Pereira, A. (2015, December). An algorithmic trading agent based on a neural network ensemble: a case of study in North American and Brazilian stock markets. In 2015 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) (Vol. 2, pp. 230-233). IEEE.
16.He, J. Y. D., & Ewbank, J. (2021). Development and Evaluation of a Machine Learning-Based Value Investing Methodology. In Advances in Data Science and Information Engineering (pp. 953-959). Springer, Cham.
17.Khaidem, L., Saha, S., & Dey, S. R. (2016). Predicting the direction of stock market prices using random forest. arXiv preprint arXiv:1605.00003.
18.Kirk, R. E. (1982). Experimental Design, Brooks. Cole Publishing Co., Wadsworth. Inc., Belmont, California, 94002, 534.
19.Lakonishok, J., Shleifer, A., & Vishny, R. W. (1994). Contrarian investment, extrapolation, and risk. The journal of finance, 49(5), 1541-1578.
20.Levine, T. R., & Hullett, C. R. (2002). Eta squared, partial eta squared, and misreporting of effect size in communication research. Human Communication Research, 28(4), 612-625.
21.Li, H., Yang, Z., & Li, T. (2014). Algorithmic trading strategy based on massive data mining. Stanford University Stanford.
22.Markowitz, H. M. (1952). Portfolio Selection, the journal of finance. 7 (1).
23.Ou, J. A., & Penman, S. H. (1989). Financial statement analysis and the prediction of stock returns. Journal of accounting and economics, 11(4), 295-329.
24.Ramaswami, M. (1994). Return enhancement through size and style management. Blending Quantitative and Traditional Equity Analysis (AIMR).
25.Rasekhschaffe, K. C., & Jones, R. C. (2019). Machine learning for stock selection. Financial Analysts Journal, 75(3), 70-88.
26.Stober, T. L. (1992). Summary financial statement measures and analysts' forecasts of earnings. Journal of Accounting and Economics, 15(2-3), 347-372.
27.Welch, B. L. (1938). The significance of the difference between two means when the population variances are unequal. Biometrika, 29(3/4), 350-362.