摘要
自动化交易是现代金融领域的研究热点,而交易策略是其核心。技术指标分析中,每一种指标都有其优势和劣势,单一的指标经常产生导致亏损的虚假信号。为了提高交易信号的质量和可靠性,针对外汇市场的多变性和存在诸多不确定性的客观事实,本文引入证据理论来处理不同技术指标分析方法结论存在的差异;将不同的指标作为独立的证据源,用D-S合成规则对各个指标分析方法的结果予以融合,建立了基于证据理论的多指标融合外汇交易模型,给出了基于证据理论的交易框架。根据技术指标的特点及交易原理,构造了指标证据的基本概率分配函数。最后,通过实例分析验证了该方法的科学性和有效性。
Automated Trading is a hot topic in the field of modern financial research, and trading strategy is its core. Each indicator has its advantage and disadvantage, single indicator often produces false signal which may cause loss. In order to improve quality and reliability of trading signal, for too many uncertainties exist in the foreign exchange market and the foreign exchange market itself changes constantly, the evidence theory is introduced to deal with the difference of different indicators analysis methods in the paper. Different indicators are regards as independent evidence resources and the results of different indicators analysis methods are fused with the Dempster-Shafer combination rule. The trading model of multiple indicators fusion based on evidence theory is created, the framework of trading strategy based on evidence theory is proposed. The basic probability assignment functions of indicators evidence are constructed according to the characteristics and the trading principles of technical indictors. Finally, an empirical analysis shows that the method is scientific and efficient.
出处
《数理统计与管理》
CSSCI
北大核心
2013年第3期452-461,共10页
Journal of Applied Statistics and Management
基金
国家自然科学基金创新研究群体科学基金资助(70821061)
国家自然科学基金青年项目(70901003)资助
关键词
证据理论
基本概率分配函数
技术指标
交易策略
evidence theory, basic probability assignment function, technical indicator, trading strategy
