DOI:
关键词:
摘要: Mostly used estimators of Hurst exponent for detection long-range dependence are biased by presence short-range in the underlying time series. We present confidence intervals estimates rescaled range and modified range. show that difference expected values enables us to use both methods together clearly distinguish between two types processes. The further applied on Dow Jones Industrial Average 1944 2009 returns do not any whereas volatility shows process.