摘要
文章研究了中国大连商品交易所大豆期货连续合约1994-2003年收益时间序列,并以该序列2003年第一个样本数据为分界点,建立了两子序列,分别进行了统计学分析,发现两子序列分布均是非正态的,较正态分布有尖峰厚尾的特征,具有记忆效应。并且,进一步根据两子序列的波动集群性建立一系列GARCH模型,对中国大豆期货的两个收益序列的波动性进行分析,并比较了二者的异同。
This article studied the return of soybean futures contract from 1994 to 2003 in Dalian Commodity Exchange market. The author splits the time series into two sub-series, with the first sample datum in 2003 as the dividing point, and conduce statistical analysis. The author founds that the two sub-series were non-normally distributed, and exhibited higher kurtosis and thicker tails, and has a memory effect. Moreover, the author sets a series of GARCH model to describe the volatility clustering of the two return series, analyzed the volatility of the return of soybean futures contracts in our country, and made a comparison towards the similarities and differences.
出处
《统计与信息论坛》
2005年第5期102-106,共5页
Journal of Statistics and Information
