NONLINEAR PROPERTIES, FRACTAL BEHAVIOR AND LONG-RANGE CORRELATION ANALYSIS OF THE CHINESE FUND MARKET
Abstract
Applying the statistical hypothesis testing, we investigate several nonlinear properties embedded in the return series of the Chinese Fund Market (CFM), which suggests the series is non-normal, auto-correlative and heteroskedastic. We hereby analyze the Hurst exponent of the return series in different timescales on the basis of the detrended fluctuation analysis (DFA) algorithm, and discuss the fractal behavior of the CFM. Furthermore, by studying the correlation of different weights in the volatility, we find the persistent long-range power-law correlation exists in the time series. We also provide hints that the above statistical properties are insensitive to the funds kind, and may be irrelevant to the market phases. Our work may reveal the self-similarity characteristics of the financial market and show a better understanding of the CFM.
You currently do not have access to the full text article. |
---|