基于GNP方法的外汇市场交易策略研究

发布时间：2018-02-16 22:38

本文关键词： GNP-RL 交易算法外汇市场交易策略　出处：《管理现代化》2014年06期 　论文类型：期刊论文

【摘要】：通过一种全新的进化算法GNP-RL(Genetic Network Programming with Reinforcement Learning)来构建外汇市场交易模型。作为一种图状的进化算法,GNP已经被成功运用到多种动态的环境中。首次将GNP方法运用到外汇市场,并通过对算法产生的超额收益验证外汇市场的有效性。实证结果表明:在使用样本外数据测试的情况下,基于GNP的交易策略是可以产生超额收益的;这种现象可以用适应性市场假说(AMH)来说明。
[Abstract]:A new evolutionary algorithm, GNP-RL(Genetic Network Programming with Reinforcement learning, is used to construct a foreign exchange market model. As a graphic evolutionary algorithm, GNP-RL(Genetic has been successfully applied to many dynamic environments. GNP method has been applied to foreign exchange market for the first time. The validity of the foreign exchange market is verified by the excess return generated by the algorithm. The empirical results show that the trading strategy based on GNP can produce excess return under the condition of using the data tested outside the sample. This phenomenon can be explained by the adaptive market hypothesis (AMH).
【作者单位】：上海财经大学统计与管理学院;
【基金】：国家自然科学基金青年科学基金项目(71101083);国家自然科学基金重点项目(71331006)
【分类号】：F832.52;F224

【参考文献】