基于自动化交易平台的高频交易及统计套利分析和研究

发布时间：2018-02-26 00:16

本文关键词： 高频交易统计套利配对交易协整套利 BP模型　出处：《复旦大学》2014年硕士论文　论文类型：学位论文

【摘要】：数量化金融理论的发展,极大的刺激和体现了信息技术在金融领域日益卓越的重要性。不管是基于事务性的数据存储处理还是基于交易性的订单和风险管理,巨大的数据存储的需求和对时间实时性的要求,计算机都扮演了极其重要和不可替代的角色。高频交易出现及其核心是从金融高频序列数理模式转换过来的策略自适应算法,依靠金融市场交易信息的滴答数据不断输入,通过概率先验判断及其他数理统计模型的推理和复杂数学公式一起组成相应的计算机量化模型模式进行金融市场未来走势的概率性判断。统计套利的出现则提供了根据这种概率先验判断和数理统计模型的理论依据来进行走势判断的可能性。它利用在市场行情数据中寻找统计学上一定置信区间下可信的交易,用承担较少和可控的风险来获得较大收益。本文主要用具体的实例对统计套利和高频交易结合交易进行具体数据的研究和论证,侧重于中国证券市场的具体股票案例。根据相关的理论依据对统计套利和其在高频交易下的案例以及数据挖掘下进行全面的介绍、研究和绩效评估。全文的主要部分的相关内容：引言主要简单介绍此论文的研究目的、选题意义、文献综述、研究方法和创新之处。第一章主要介绍金融交易的相关理论,包括交易的撮合形式,交易的策略分类和交易频率的类别。第二章开始介绍统计套利的理论和策略,包括其理论知识,符合统计套利的先决条件和其相关传统套利模型。从简单的配对交易开始,到稍微复杂的均值回归模型,多因子模型和协整套利模型等。最后发展到基于数据挖掘技术理论和应用包括神经网络、遗传算法和混沌时间序列在内的统计套利的分析和研究。第三章进行统计套利的实证分析,这是本文的核心部分。选取相应的中国证券市场的股票案例对配对模型和协整模型的统计套利进行实例性的论证分析和研究,并进一步研究协整模型的统计套利在高频交易下的实例分析,最后论证数据挖掘中BP模型的案例验证。末章进行关于高频统计套利自动化交易平台结构介绍和模型测试,对其进行绩效评估方法的研究及展望,并结合资本市场上实际发生的光大证券乌龙指事件进行分析和阐述。
[Abstract]:The development of quantitative financial theory has greatly stimulated and demonstrated the increasing importance of information technology in the financial field, whether it is based on transactional data storage processing or transactional order and risk management. Because of the huge demand of data storage and the requirement of real-time time, computer plays an important and irreplaceable role. The emergence of high-frequency trading and its core is a self-adaptive strategy algorithm, which is transformed from the mathematical mode of financial high-frequency sequence. Ticking data, which rely on financial market transaction information, The probabilistic judgment of the future trend of the financial market is carried out through probabilistic priori judgment and the reasoning of other mathematical statistical models and the complex mathematical formula. The appearance of statistical arbitrage provides the basis for the probability judgment of the future trend of the financial market. The possibility of judging the trend according to the theoretical basis of the probabilistic priori judgment and the mathematical statistical model is given. It uses the market price data to find credible transactions with a statistical confidence interval. This paper mainly uses the concrete examples to study and prove the specific data of statistical arbitrage and high-frequency trading. Focusing on the specific stock cases in China's securities market. According to the relevant theoretical basis to carry on the comprehensive introduction to the statistical arbitrage and its cases under high-frequency trading and data mining. Research and performance evaluation. The main part of the full text related content: the introduction mainly introduces the purpose of this paper, the significance of the topic, literature review, research methods and innovations. The first chapter mainly introduces the relevant theories of financial transactions. The second chapter introduces the theory and strategy of statistical arbitrage, including its theoretical knowledge. Consistent with the prerequisites for statistical arbitrage and its associated traditional arbitrage models. From simple pairing transactions to slightly more complex mean regression models, Finally, based on the theory and application of data mining technology, including neural network, The analysis and research of statistical arbitrage including genetic algorithm and chaotic time series. This is the core of this paper. Select the corresponding stock case of Chinese stock market to analyze and study the statistical arbitrage of pairing model and cointegration model. And further study the statistical arbitrage of cointegration model in high-frequency trading case analysis, and finally demonstrate the data mining BP model case verification. The last chapter introduces the structure of high-frequency statistical arbitrage automated trading platform and model testing. The research and prospect of its performance evaluation method, and the analysis and exposition of the Oolong finger event of Everbright Securities in the capital market.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.51

【相似文献】