金融复杂系统的特征研究及其交易策略构建

发布时间：2018-01-17 23:32

本文关键词：金融复杂系统的特征研究及其交易策略构建　出处：《宁波大学》2014年硕士论文　论文类型：学位论文

更多相关文章： 标度指数 R/S分析 DFA分析 小波分析

【摘要】：复杂性是客观事物不同层次的跨越，存在于客观事物不同层次之间，金融市场是一个复杂的系统，在研究金融系统的复杂性时我们一般从下面几个基本特征研究：证券价格序列的峰度和偏度、收益率的概率分布以及经验分析等等。我国证券市场股票收益率具有典型的“尖峰厚尾”分布，所以在分析收益率的概率分布时，我们采用了列维平稳分布，因为列维分布跟正态分布的区别是列维分布不仅具有对称性而且还具有尖峰厚尾特征。经验分析中我们采用对数回复率进行分析，分析上证指数对数回复率的自相自相关性。基于对金融市场的长程记忆性研究，本文利用重标极差分析(R/S)和降趋脉动分析(DFA)这两种方法分别研究了我国上证指数的长程记忆性。证实了中国证券市场上存在弱式长程记忆，以及存在多标度特征现象。本文根据中国证券市场上的弱式长程记忆性，以及它的物理意义，利用以往历史数据，通过R程序构建了模型回测系统，检验是否可以根据弱式记忆性进行投资。通过对模型的收益率和大盘指数的收益率比较，发现通过弱式长程记忆可以帮助投资者做投资决策。证券市场上的价格时间序列往往存在噪声，噪声严重的干扰了投资者对趋势的判断。而数学上的小波分析可以通过分解降噪，平滑信号，通过分解处理的信号，然后重新构造新的信号，新的信号往往平滑了很多。利用这个原理，本文对上证指数价格序列做了小波降噪分析，降噪处理后的时间序列的确平滑了很多。经过降噪处理后的时间序列，基本上可以认为是平稳序列。本文选取了自回归模型(AR)，，对经过处理后的平稳时间序列进行预测，并且对预测结果进行了分析。
[Abstract]:Complexity is the leapfrogging of different levels of objective things and exists between different levels of objective things. Financial market is a complex system. In studying the complexity of financial system, we generally study the following basic characteristics: kurtosis and skewness of securities price series. The probability distribution and empirical analysis of the return rate. The stock return rate of China's stock market has a typical "peak thick tail" distribution, so in the analysis of the probability distribution of the return rate, we use Levi stationary distribution. Because the difference between Levi distribution and normal distribution is that the Levi distribution not only has symmetry but also has the characteristic of peak and thick tail. This paper analyzes the autocorrelation of logarithmic recovery rate of Shanghai Stock Exchange Index. Based on the long-term memory of financial markets. In this paper, we study the long range memory of Shanghai stock index by using the method of rescaled range analysis (R / S) and DFAA. It is proved that there is a weak long term memory in Chinese stock market. According to the weak long range memory in Chinese stock market and its physical meaning, this paper constructs a model retrieval system by using the historical data and R program. By comparing the return rate of the model with that of the large market index, it is found that weak long-term memory can help investors to make investment decisions. There is often noise in the time series of price in the stock market, which seriously interferes with the investors' judgment of the trend. The wavelet analysis in mathematics can smooth the signal by decomposing the noise. By decomposing the processed signal and then reconstructing the new signal, the new signal is often smooth. Using this principle, this paper does wavelet denoising analysis to the price sequence of Shanghai stock index. After noise reduction, the time series are smooth. After the noise reduction, the time series can be considered as stationary series. In this paper, the autoregressive model is selected. The stationary time series after processing are predicted, and the prediction results are analyzed.
【学位授予单位】：宁波大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F830.9

【参考文献】