论文标题
全球股票市场的动态特征基于时间依赖的tsallis非扩展统计数据和广义HURST指数
Dynamical Characteristics of Global Stock Markets Based on Time Dependent Tsallis Non-Extensive Statistics and Generalized Hurst Exponents
论文作者
论文摘要
我们使用时间依赖的扩展Tsallis统计数据对股票市场指数进行非线性分析。具体而言，我们在特定时间段内评估了Q-triplet，目的是证明基础市场动态的扩展特征的时间依赖性。我们对四个主要全球市场的每日近距离时间表进行分析（标准普尔500，东京 - 尼克基，法兰克福 - 戴克斯，伦敦LSE）。为了进行比较，我们还使用GHE方法来计算时间依赖性的广义HURST指数（GHE）HQ，从而估算了索引动力学的多标准特征的时间演化。 We focus on periods before and after critical market events such as stock market bubbles (2000 dot.com bubble, Japanese 1990 bubble, 2008 US real estate crisis) and find that the temporal trends of q-triplet values significantly differ among these periods indicating that in the rising period before a bubble break, the underlying extended statistics of the market dynamics strongly deviates from purely stochastic behavior, whereas, after the breakdown, it gradually converges to the高斯式行为，这是高效市场的特征。我们还得出结论，Tsallis Q-Triplet的相对时间变化模式可以连接到市场动态的不同方面，并揭示有关市场状况的有用信息，尤其是那些股票市场泡沫发展的市场。我们发现了Q-三重方案中指数相对变化的特定时间模式和趋势，这些变化在库存市场气泡破裂之前和之后区分了时期。内源性股市和外源股市危机之间的差异也被Tsallis Q-triplet的时间变化所捕获。最后，我们引入了两个新的时间依赖性经验指标（Q-Metrics），它们是Tsallis Q-Triplet的功能。
We perform non-linear analysis on stock market indices using time-dependent extended Tsallis statistics. Specifically, we evaluate the q-triplet for particular time periods with the purpose of demonstrating the temporal dependence of the extended characteristics of the underlying market dynamics. We apply the analysis on daily close price timeseries of four major global markets (S&P 500, Tokyo-NIKKEI, Frankfurt-DAX, London-LSE). For comparison, we also compute time-dependent Generalized Hurst Exponents (GHE) Hq using the GHE method, thus estimating the temporal evolution of the multiscaling characteristics of the index dynamics. We focus on periods before and after critical market events such as stock market bubbles (2000 dot.com bubble, Japanese 1990 bubble, 2008 US real estate crisis) and find that the temporal trends of q-triplet values significantly differ among these periods indicating that in the rising period before a bubble break, the underlying extended statistics of the market dynamics strongly deviates from purely stochastic behavior, whereas, after the breakdown, it gradually converges to the Gaussian-like behavior which is a characteristic of an efficient market. We also conclude that relative temporal variation patterns of the Tsallis q-triplet can be connected to different aspects of market dynamics and reveals useful information about market conditions especially those underlying the development of a stock market bubble. We found specific temporal patterns and trends in the relative variation of the indices in the q-triplet that distinguish periods just before and just after a stock-market bubble break. Differences between endogenous and exogenous stock market crises are also captured by the temporal changes in the Tsallis q-triplet. Finally, we introduce two new time-dependent empirical metrics (Q-metrics) that are functions of the Tsallis q-triplet.
