针对非参数方法研究国内股市长记忆性时结论参差不齐的现状,本文研究了更为稳健的半参数估计方法,即局部W h ittle(LW)估计和对数周期图(LP)回归。通过对不同频率高频数据的分析,证实了LW估计方法尽管需要数值最优化,但仍然要优于LP回...针对非参数方法研究国内股市长记忆性时结论参差不齐的现状,本文研究了更为稳健的半参数估计方法,即局部W h ittle(LW)估计和对数周期图(LP)回归。通过对不同频率高频数据的分析,证实了LW估计方法尽管需要数值最优化,但仍然要优于LP回归。进而将LW估计首次应用于中国股市,结果表明不同频率绝对收益序列的长记忆强度基本一致;同时发现,重大突发事件发生时的长记忆性表现得最为强烈,且事件后比事件前表现的要强烈,这说明股票市场的溢出效应在事件后增强,此项结论对我国证券市场有一定的借鉴意义。展开更多
Let (X,Y) be an R^d×R^1 valued random vector (X_1,Y_1),…, (X_n,Y_n) be a random sample drawn from (X,Y), and let E|Y|<∞. The regression function m(x)=E(Y|X=x) for x∈R^d is estimated by where, and h_n is a p...Let (X,Y) be an R^d×R^1 valued random vector (X_1,Y_1),…, (X_n,Y_n) be a random sample drawn from (X,Y), and let E|Y|<∞. The regression function m(x)=E(Y|X=x) for x∈R^d is estimated by where, and h_n is a positive number depending upon n only, nad K is a given nonnegative function on R^d. In the paper, we study the L_p convergence rate of kernel estimate m_n(x) of m(x) in suitable condition, and improve and extend the results of Wei Lansheng.展开更多
文摘针对非参数方法研究国内股市长记忆性时结论参差不齐的现状,本文研究了更为稳健的半参数估计方法,即局部W h ittle(LW)估计和对数周期图(LP)回归。通过对不同频率高频数据的分析,证实了LW估计方法尽管需要数值最优化,但仍然要优于LP回归。进而将LW估计首次应用于中国股市,结果表明不同频率绝对收益序列的长记忆强度基本一致;同时发现,重大突发事件发生时的长记忆性表现得最为强烈,且事件后比事件前表现的要强烈,这说明股票市场的溢出效应在事件后增强,此项结论对我国证券市场有一定的借鉴意义。
文摘Let (X,Y) be an R^d×R^1 valued random vector (X_1,Y_1),…, (X_n,Y_n) be a random sample drawn from (X,Y), and let E|Y|<∞. The regression function m(x)=E(Y|X=x) for x∈R^d is estimated by where, and h_n is a positive number depending upon n only, nad K is a given nonnegative function on R^d. In the paper, we study the L_p convergence rate of kernel estimate m_n(x) of m(x) in suitable condition, and improve and extend the results of Wei Lansheng.
