政大機構典藏-National Chengchi University Institutional Repository(NCCUR):Item 140.119/130991

政大機構典藏-National Chengchi University Institutional Repository(NCCUR):Item 140.119/130991

English | 正體中文 | 简体中文 | Post-Print筆數 : 27 | Items with full text/Total items : 113822/144841 (79%)
Visitors : 51819122 Online Users : 504

RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.

Scope

please add "double quotation mark" for query phrases to get precise results

please goto advance search for comprehansive author search

Adv. Search

Home ‧ Login ‧ Upload ‧ Help ‧ About ‧ Administer

Goto mobile version

政大典藏 > College of Commerce > Department of Money and Banking > Theses > Item 140.119/130991

Please use this identifier to cite or link to this item: https://nccur.lib.nccu.edu.tw/handle/140.119/130991

Title:	資產配置基於集成學習的多因子模型－以台灣股市為例 Asset Allocation Based on Ensemble-Learning Assisted Multi-Factor Models– Taiwan Stock Market as an Example
Authors:	陳昱安 Chen, Yu-An
Contributors:	江彌修 Chiang, Mi-Hsiu 陳昱安 Chen, Yu-An
Keywords:	因子選股機器學習集成學習隨機森林台股市場 Factors stock selection Machine learning Ensemble-learning XGBoost Random forest Taiwan stock market
Date:	2020
Issue Date:	2020-08-03 17:38:21 (UTC+8)
Abstract:	本研究使用了三種傳統多因子選股模型以及結合了因子選股的兩種集成學習法為基礎之機器學習分類模型Extreme Gradient Boosting (XGBoost)、Random Forest來建構選股模型，並且比較了傳統多因子選股模型以及機器學習因子選股模型之策略績效，同時觀察兩種機器學習模型之間預測效果以及策略績效的差異性。而本研究所採用之資產標的為台灣股票市場之上市股票，樣本回測期間採自2010/1/1 至2020/1/1之所有台灣上市股票，因子特徵選取方面撇除掉過去研究常用之基本面數據，採用價量面以及台灣市場獨有之籌碼面資料，實證結果顯示，兩種機器學習模型在樣本內回測期間大幅優於傳統多因子選股模型，而在樣本外回測期間策略績效表現亦較傳統多因子選股模型出色，顯示了機器學習模型挖掘出資產報酬趨勢的能力。仔細比較XGBoost和Random Forest的策略績效後，可以發現到XGBoost優於Random Forest，表示帶有懲罰項係數防止決策樹過度擬合之XGBoost模型在樣本外表現上優於Random Forest，完整體現了金融市場隨著時間不斷變化的特性，模型優化了過度擬合歷史數據的缺點，提升模型在樣本外的性能。 In this paper, we implement three traditional multi-factor stock selection models and two ensemble-learning models including Extreme Gradient Boosting (XGBoost) and Random Forest combined with stock factors. We compared the strategy performance of traditional factor stock selection model and ensemble-learning factor stock selection model, and the difference between the prediction effectiveness and strategy performance of the two ensemble-learning models. The assets used in this research are the listed stocks in Taiwan stock market and the backtesting period is from 2010/1/1 to 2020/1/1. The stock factors include technical data and the unique trading volume data of Taiwan stock market and exclude fundamental data that commonly used in the research before. The empirical results show that the two ensemble-learning models outperform the traditional multi-factor stock selection model during the whole period. This results prove the ability of ensemble-learning models to capture asset return trend. In the results, we can also find that XGBoost outperforms Random Forest, indicating the model with penalty coefficients to prevent over-fitting outperforms, fully reflecting the time-varying of financial market, the model optimizes the disadvantage in over-fitting historical data and improves the performance of the model outside the sample.
Reference:	[1] Ahmad I., Basheri M., Iqbal M J., Raheem A. (2017). Improving Factor-Based Quantitative Investing by Forecasting company fundamentals. [2] Boivin J., S Ng. (2006). Are more data always better for factor analysis. Journal of Econometrics. 169-194. [3] Breiman L., J Friedman., C Stone. (1984). Classification and regression Trees. Wadsworth and Brooks Monetarey. [4] Breiman L. (2001). Random Forests. Machine Learning. 5-32. [5] Cao L., Tay F H. (2003). Support vector machine with adaptive parameters in financial time series forecasting. Neural Netw. [6] Deroski S., Enko B. (2004). Is combining classifiers with stacking better than selecting the best one? Machine Learning. 255-273. [7] Fama Eugene F., Kenneth R French. (1992). The cross-section of expected stock returns. The Journal of Political Economy. [8] Fan A., Palaniswami M. (2001). Stock selection using support vector machines. In proceedings of the international joint conference on Neural Networks. [9] Liaw A., Wiener M. (2002). Classification and regression by randomForest. R news. 18-22. [10] Ruder S. (2016). An overview of gradient descent optimization algorithms. arXiv preprint arXiv:1609.04747. [11] Sharpe, W.F. (1970). Portfolio theory and capital markets. The Journal of Finance, 1-18. [12] Sharpe, W.F. (1994). The sharpe ratio. The Journal of Portfolio Management, 49-58. [13] Van der Hart J., Slagter E., Van Dijk D. (2003). Stock selection strategies in emerging markets. The Journal of Empirical Finance. 105-132. [14] Yangming Z., Guoping Q. (2018). Random forest for label ranking. 99-109. [15] Zhu M., Philpotts D., Sparks R., Stevenson M J. (2011). A hybrid approach to combining CART and logistic regression for stock ranking. The Journal of Portfolio Management. 100-109. [16] Zhu M., Philpotts D., Stevenson M J. (2012). The benefits of tree-based models for stock selection. 437-448. [17] Zheng T., Ziqin Y., Guangwei Z. (2019). Stock selection with random forest: An exploitation of excess return in the Chinese stock market. Heliyon.
Description:	碩士國立政治大學金融學系 107352021
Source URI:	http://thesis.lib.nccu.edu.tw/record/#G0107352021
Data Type:	thesis
DOI:	10.6814/NCCU202000702
Appears in Collections:	[Department of Money and Banking] Theses

Files in This Item:

File	Description	Size	Format
202101.pdf		2892Kb	Adobe PDF2	0	View/Open

All items in 政大典藏 are protected by copyright, with all rights reserved.

社群 sharing

著作權政策宣告 Copyright Announcement

1.本網站之數位內容為國立政治大學所收錄之機構典藏，無償提供學術研究與公眾教育等公益性使用，惟仍請適度，合理使用本網站之內容，以尊重著作權人之權益。商業上之利用，則請先取得著作權人之授權。
The digital content of this website is part of National Chengchi University Institutional Repository. It provides free access to academic research and public education for non-commercial use. Please utilize it in a proper and reasonable manner and respect the rights of copyright owners. For commercial use, please obtain authorization from the copyright owner in advance.

2.本網站之製作，已盡力防止侵害著作權人之權益，如仍發現本網站之數位內容有侵害著作權人權益情事者，請權利人通知本網站維護人員(nccur@nccu.edu.tw)，維護人員將立即採取移除該數位著作等補救措施。
NCCU Institutional Repository is made to protect the interests of copyright owners. If you believe that any material on the website infringes copyright, please contact our staff(nccur@nccu.edu.tw). We will remove the work from the repository and investigate your claim.

DSpace Software Copyright © 2002-2004 MIT & Hewlett-Packard / Enhanced by NTU Library IR team Copyright © - Feedback