基于分形理论的上证股指非线性特征研究

VIP免费
3.0 陈辉 2024-11-19 5 4 1.11MB 77 页 15积分
侵权投诉
i
对资本市场诸多问题的研究,如资本资产定价、金融风险的测度与防范、证
券投资组合的选择,以及金融衍生品定价等问题,都要首先依赖于对资本市场基
本均衡特性和波动特性的认识与描述。资本市场的本质是非线性复杂系统,对其
行为特征的探讨需要打破传统的线性范式,结合非线性理论进行分析。
分形理论是非线性系统理论的重要组成部分,其诞生对人们的自然观、科学
观、方法论和思维方式产生了深远的影响,同时也为金融市场研究提供了新的工
具,开辟了新的视野。本文将该理论应用于上证股票市场,按检验、分析、应用
的三大结构层次,从指数收益序列非线性结构的统计检验出发,拒绝了收益率呈
正态分布的假设,BDS 检验方法验证了股指非线性结构的存在,随后分别基于
序列数据与股指图像设计编程方案,在两种算法下获得维数参量,提供了盒维的
实际计算方法,对股指标度不变性和复杂程度进行了描述。再者,根据我国沪市
股票市场的平均循环长度,在对重标极差分析法做出周期分段改进的基础上,求
出描述非线性时间序列长期相关性的重要指标 Hurst 指数,并对指数的有效性进
行了检验,得出我国上证股票市场存在长期记忆性的结论。最后结合数据,以沪
12 支银行股票为例讨论了非线性特征参数在股票风险识别中的启发性应用,
开创了滚动周期窗口下的 Hurst 指数计算法,通过与历史大盘走势的比较,对市
场趋势预测和预警方法进行了拓展性思考。
文章在分形视角下较全面地对上证股指非线性结构进行了探讨,在具体问题
的分析计算和应用上含有一定特色,研究内容和所得结论对于突破传统理论束缚、
在整体上科学地认识和把握我国股票市场非线性特征、准确计算非线性参数指标、
给予投资者正确的决策分析和合理建议具有重要意义,并拓展了风险管理和股价
预测研究的新思路,为后续一系列重要金融问题的优化和改进方法提供了理论参
考价值。
关键词:分形 股票指数 非线性检验 盒维 长期相关性 风险管理
ii
ABSTRACT
For many researches of capital market problems, such as capital asset pricing,
financial risk measure and prevention, the choice of portfolio investment, as well as
financial derivatives pricing etc, the premise lies on the understanding and description
of market’s equilibrious and fluctuant character . The capital market is a non-linear and
complex system, that nature dooms its behavior traits should be found and analyzed
depending on non-linear theory rather than traditional linear paradigm.
Fractal theory is an important component of nonlinear systems theoretics, whose
birth not only made a profound impact that changes the human being’s view of nature,
scientific concept, methodology and ways of thinking, but also provided a new tool for
financial research, even opened up a new horizon. According to a three-level scheme
formed from examination, analysis and application, this thesis applies this theory into
the reality of Shanghai stock market. With non-linear statistical tests of index sequence,
the original assumption of yield normal distribution is rejected and the existence of
nonlinear structure is proved by BDS test. Then two algorithms, stock data-based and
image-based respectively, are programmed to compute the stock dimension. That
provides a practical calculation method for box dimension to describe scale invariance
and complexity feature of stock index. Besides, in the light of the average cycle length
of Shanghai stock market, sub-section amelioration for the rescaling range analysis is
made to obtain Hurst exponent, which is a very important parameter to indicate the
long-term relevance of nonlinear time series. Finally, combined with recent data,
taking 12 bank stocks in Shanghai market as an example, a revelatory application of
non-linear characteristic parameters for stock risk identification is discussed.
Afterwards, comparing with the real history trend and facts of stock market, an
innovative Hurst index calculation method under rolling cycle window is created and
put into demonstration. This method can be seemed as an initial thinking for the market
forecasting and early warning.
From fractal theory perspective, the paper roundly probes into the non-linear
structure of Shanghai stock index with some specific unique features in both analysis
computing and applications. The content and conclusions of this study make great
significance in many aspects as breaking out the constraints of traditional theory,
understanding and grasping the non-linear characteristics of China's stock market truly
iii
in a scientific way, calculating the non-linear parameters accurately, giving investors
proper decision-making analysis and reasonable proposals. In addition, it also ignites
new ideas for risk management and trend forecasting. And further more, the research
will supply influential theoretical reference value for the follow-up optimization and
improvement of a series crucial financial issues.
Key Words: fractal, stock index, non-linear test, box dimension,
long-term dependence, risk management
iv
中文摘要
ABSTRACT
第一章 .................................................................................................................. 1
§1.1 研究背景与意义 ............................................................................................... 1
§1.1.1 背景 ....................................................................................................... 1
§1.1.2 意义 ....................................................................................................... 1
§1.2 资本市场的非线性理论研究 ........................................................................... 2
§1.3 分形理论在资本市场中的发展 ....................................................................... 7
§1.4 研究思路与框架 ............................................................................................. 10
第二章 基本概念与理论基础 ...................................................................................... 13
§2.1 分形与维数 ...................................................................................................... 13
§2.1.1 分形 ...................................................................................................... 13
§2.1.2 维数 ...................................................................................................... 14
§2.2 分数布朗运动与Hurst指数 ............................................................................. 18
§2.3 重标极差R/S分析 ........................................................................................... 20
§2.3.1 R/S定义 ................................................................................................ 21
§2.3.2 Hurst指数的检验 ................................................................................. 22
§2.3.3 平均循环周期 ..................................................................................... 23
§2.4 本章小结 ......................................................................................................... 24
第三章 上证综指非线性结构检验 .............................................................................. 25
§3.1 数据采集与处理 .............................................................................................. 25
§3.2 正态性经验 ..................................................................................................... 26
§3.2.1 直观检验 ............................................................................................. 26
§3.2.2 统计量检验 ......................................................................................... 28
§3.3 非线性BDS检验 ............................................................................................. 30
§3.3.1 BDS统计原理 ...................................................................................... 30
§3.3.2 Q统计量与序列线性成分剔除 ........................................................... 31
§3.3.3 BDS检验结果 ...................................................................................... 34
§3.4 本章小结 ......................................................................................................... 35
v
第四章 上证综指非线性分形特征研究 ...................................................................... 37
§4.1 股指分形结构的直观表现 ............................................................................. 37
§4.2 股指分形维(盒维)计算 .............................................................................. 38
§4.1.1 基于股指数据的盒维测量 ................................................................. 38
§4.2.2 基于股指图像的盒维算法 ................................................................. 42
§4.3 股指长记忆过程与周期性研究 ...................................................................... 45
§4.3.1 R/S算法 ................................................................................................ 45
§4.3.2 R/S实证分析 ........................................................................................ 46
§4.4 本章小结 ......................................................................................................... 51
第五章 理论实证与应用研究 ...................................................................................... 53
§5.1 沪市银行个股风险分析 ................................................................................. 53
§5.2 滚动周期窗口的Hurst方法研究 .................................................................... 57
§5.3 本章小结 ......................................................................................................... 61
第六章 总结与展望 ...................................................................................................... 63
§6.1 论文总结 ......................................................................................................... 63
§6.2 后续研究展望 ................................................................................................. 64
参考文献 ........................................................................................................................ 67
在读期间公开发表的论文和承担科研项目及取得成果 ............................................ 73
............................................................................................................................ 75
第一章
1
第一章
§1.1 研究背景与意义
§1.1.1 背景
中国股市自建立以来的短短18年中,不断以惊人速度发展着。据统计,截至
20081212日,沪深两市共有A股账户1.2亿户,B股账户240.15万户。中国股票
市场不仅吸引了亿万人民加入股民大军,同时也吸引了大量学者对该问题进行广
泛探讨与研究。
一方面,股票市场作为资源配置的手段,在资本再配置环节中起着重要的作
用。股指态势不仅反映了国家经济的宏观发展状况,更与亿万百姓的切身利益息
息相关,关乎国计民生。伴随着近年来我国股票市场规模的不断扩大,交易手段
的日渐成熟,对市场真实特征进行有效论述的要求越加迫切,对其发展趋势的合
理描述方法成为人们关注的热点问题。
另一方面,当下,来势汹汹的金融海啸席卷全球。2008年以来,受到美国次
贷危机影响,国内股票市场经历了大幅调整,沪深两市总市值缩水一半以上,08
下半年沪指自突破6000点大关后回落降至1700余点,这一由“牛”转“熊”的大
幅跳水造成市场萎靡状态持续至今,广大股民无不“谈股色变。股市摩天轮的疯
狂让中国股民在体会到切肤之痛后真正明白了“股市有风险,投资需谨慎”这句
话的深刻含义。高风险意味着高回报,“危”“机”并存,股海沉浮中既有“万
贯家财化乌有”的盲目入市者,也有成就“股神巴菲特神话”的获益投资人,关
键在于是否能认识到股市真实的行为特征。伴随着股市风险,股民的投资理性与
谨慎性也在不断提高,金融风险衡量与趋势预测方法日益受到重视。在现阶段,
从历史数据中挖掘股指特征,分析股票风险与运行趋势,为投资决策提供适当的
理论指导和支持是必要的。
§1.1.2 意义
金融市场(如股票市场)是人类社会的产物,但至今人类对其运行规律仍懵
懂不明,不断探索着新的市场行为解释方法。传统金融普遍建立在有效市场假说
Efficient Market Hypothesis, EMH)之上,假定市场无磨擦、无交易成本、无信
摘要:

i摘要对资本市场诸多问题的研究,如资本资产定价、金融风险的测度与防范、证券投资组合的选择,以及金融衍生品定价等问题,都要首先依赖于对资本市场基本均衡特性和波动特性的认识与描述。资本市场的本质是非线性复杂系统,对其行为特征的探讨需要打破传统的线性范式,结合非线性理论进行分析。分形理论是非线性系统理论的重要组成部分,其诞生对人们的自然观、科学观、方法论和思维方式产生了深远的影响,同时也为金融市场研究提供了新的工具,开辟了新的视野。本文将该理论应用于上证股票市场,按检验、分析、应用的三大结构层次,从指数收益序列非线性结构的统计检验出发,拒绝了收益率呈正态分布的假设,以BDS检验方法验证了股指非线性结...

展开>> 收起<<
基于分形理论的上证股指非线性特征研究.pdf

共77页,预览8页

还剩页未读, 继续阅读

作者:陈辉 分类:高等教育资料 价格:15积分 属性:77 页 大小:1.11MB 格式:PDF 时间:2024-11-19

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 77
客服
关注