Font Size: a A A

Research On Fractal Feature Of Copper Future Market In SHFE

Posted on:2015-01-25Degree:MasterType:Thesis
Country:ChinaCandidate:J Y CuiFull Text:PDF
GTID:2269330428972657Subject:Quantitative Economics
Abstract/Summary:PDF Full Text Request
By applying autocorrelation functions and GARCH model methods, we conduct empirical analysis to logarithmic return time series of copper future contract in Shanghai Future Exchange (SHFE) when time-scale τ is1,5,22,66. These return time series all exist volatility clustering phenomena and show fractal feature under various time-scales. We also introduce an index through moving-window approach to quantitatively measure the volatility clustering degree in these time series, finding that the degree of volatility clustering increases with the growth of time range τ and moving-window size.Based on the R/S analysis and the modified Rescaled Range Analysis method (Modified R/S), this paper conducts the long-term memory characteristics on various time-scale logarithmic return time series of copper future contract in Shanghai Future Exchange (SHFE). The results indicate that copper future contract market doesn’t exhibit obvious permanent trends and the price doesn’t have long-term memory by Modified R/S method, while the classic R/S method shows memory characteristics on copper future contract in SHFE.Based on the DFA analysis, this paper conducted research on various time-scale logarithmic return time series of copper future contract in Shanghai Future Exchange (SHFE) about the long-term memory characteristics and analyzed the scaling properties. The results indicate that the scaling index becomes larger and long-term memory gets strengthened with the increase of the time scaling. By applying high-level DFA analysis, it is also found that the scaling exponents don’t vary obviously with the scaling.Based on the MF-DFA analysis, this paper conducted empirical research on various time-scale logarithmic return time series of copper future contract in Shanghai Future Exchange (SHFE) and found the presence of significant Multi-fractal properties. Though shuffling procedure and phase randomization procedure, we compared the original time series and the transformed time series and analyzed thoroughly the resource of the Multi-fractal properties. The result indicates that the long-term memory and fat-tailed probability distribution contribute to the multi-fractal behavior, and long-term memory of price volatility is the primary source.
Keywords/Search Tags:Copper Future Contract, Fractal, Volatility Clustering, Long-term Memory, Self-similarity
PDF Full Text Request
Related items