摘要
以指数分布为例,提出用混合分布去拟合一组数据,比用单一的指数分布要好.一般的原始分布只对中间部分拟合较好,而对尾部拟合较差,可用混合分布来克服此缺陷.
Taking exponential distribution as example,we sum up that to fit a set of data with mixed distribution is better than with single exponential distribution. Initial distribution often fits well in the central part,but it is not the case in the tail. In this paper,the shortcoming is overcome by using the mixed distribution.
出处
《黄冈师范学院学报》
2002年第3期4-7,共4页
Journal of Huanggang Normal University