��ժҪ��Ϣ

Ӧ�ø��ͳ�� 2012, 28(3) 277-284 DOI: ISSN: 1001-4268 CN: 31-1256

��Ŀ¼ | ��Ŀ¼ | �� | �߼�� [��ӡ��ҳ] [�ر�]

ѧ��

	��չ��

	��Ϣ
	Supporting info
	PDF(174KB)
	[HTMLȫ��]
	�ο��[PDF]
	�ο��
	��뷴��
	�ѱ��Ƽ��
	��ҵ��
	��ù��
	��ñ��
	Email Alert
	��Ĺؼ��
	��
	PubMed

��ڶ��ͬ�ֲ��̬��ֺ�β��ʵ�ע

�ν��,лͥ��

�й��ѧԺ��ѧϵ

ժҪ��

�� $\100dpi \inline $\{X,X_{n},n\geq1\}$$ �Ƕ��ͬ�ֲ��̬��,
$\100dpi \inline $\ep X=0$$ �� $\100dpi \inline $\ep X^2=\sigma^{2}>0$$ , $\100dpi \inline $S_{n}=\tsm_{k=1}^{n}X_{k}$$ ,
$\100dpi \inline $\lambda(\epsilon)=\tsm_{n=1}^{\infty}\pr(|S_{n}|\geq n\epsilon)$$ .
�ڱ��, ��֤��˴�� $\100dpi \inline $C_{1}$$ �� $\100dpi \inline $C_{2}$$ ,
ʹ�ö��㹻С�� $\100dpi \inline $\epsilon>0$$ , ��в��ʽ $\100dpi \inline $C_{1}\epsilon^{3}\leq$$ $\100dpi \inline $\epsilon^{2}\lambda(\epsilon)-\sigma^{2}+{\epsilon^{2}}/{2}\leq C_{2}\epsilon^{3}$$ .

�ؼ����

A Remark on the Tail Probability Sums of i.i.d. Gaussian Random Variable

He Jianjun,Xie Tingfan

Department of Mathematics, China Jiliang University

Abstract:

Let $\100dpi \inline $\{X,X_{n},n\geq1\}$$ be a sequence
of i.i.d. Gaussian random variables with zero mean and finite
variance, and set $\100dpi \inline $S_{n}=\tsm_{k=1}^{n}X_{k}$$ ,
$\100dpi \inline $\lambda(\epsilon)=\tsm_{n=1}^{\infty}\pr(|S_{n}|\geq n\epsilon)$$ .
In this paper, we prove that there exists positive constants $\100dpi \inline $C_{1}$$ and $\100dpi \inline $C_{2}$$ ,
for small enough $\epsilon>0$, it follows that
$\100dpi \inline $\epsilon>0$$ , it follows that
$\100dpi \inline $C_{1}\epsilon^{3}\leq$$ $\100dpi \inline $\epsilon^{2}\lambda(\epsilon)-\sigma^{2}+{\epsilon^{2}}/{2}\leq C_{2}\epsilon^{3}$$ .

Keywords:

�ո�� �޻�� ��淢��

DOI:

��Ŀ:

ͨѶ��: �ν��

��߼��:

��Email:

�ο��ף�

��е��