Analysis of Fluctuation and Trend of Shenzhen Stock Market Based on ARFIMA

摘要

The empirical analysis on return series of 5-min high-frequency data of SZCI suggests that there exists the apparent evidence of long-range correlation in the two time series of G and absolute G.The absolute return appears stronger positive persistence than series G.Moreover,it shows that ARFIMA (p,d,q) models can describe properties of return fluctuation more accurately than traditional nonARFIMA (p,q) models,and ARFIMA (p,d,q) models also indicate higher fitting relative to non-ARFIMA models in predicting future return.The further study shows no significant difference between series G and series of absolute G with the apparently different fractional difference parameter d.