摘要
在线投资组合选择问题是当前量化投资领域一个重要的研究问题.为了避免剧烈波动的股票市场中过去较长时间的股价数据对当前的投资决策产生干扰,基于移动窗口设计在线投资组合策略.首先利用近期股价数据,计算所有定常再调整策略的近期表现并对其进行排序;根据其排序构造权重,对所有定常再调整策略进行加权平均,提出了基于移动窗口的策略;进一步采用适应性学习的方法选择移动窗口的长度,提出了适应性学习的策略.采用实际股价数据对提出的策略进行了实证分析,结果表明它们具有较好的性能.
Online portfolio selection is an important research problem in the field of quantitative investment. To avoid the interference with current investment decisions caused by the stock price data far from now in the intensely fluctuating stock market, online portfolio strategies based on moving window are designed. Using the recent stock price data, the recent performances of all constant rebalanced portfolios are computed and ranked. An online portfolio strategy based on moving window is designed by weighted averaging all constant rebalanced portfolios. Further using adaptive learning method to select the length of the moving window, the adaptive learning strategy is put forward. Empirical analyses are made on the proposed strategies using the real stock price data. The results show that they have better performance.
作者
杨兴雨
何锦安
沈健华
Yang Xing-yu;He Jin-an;Shen Jian-hua(School of Management, Guangdong University of Technology, Guangzhou 510520, China)
出处
《广东工业大学学报》
CAS
2018年第3期61-66,共6页
Journal of Guangdong University of Technology
基金
国家自然科学基金资助项目(71301029)
广东省高等学校优秀青年教师培养计划(Yq2013060)
广东省哲学社会科学2016年度学科共建项目(17ZS0103)
关键词
投资组合
移动窗口
适应性策略
在线算法
实证分析
portfolio
moving window
adaptive strategy
online algorithm
empirical analysis
