机构地区: 中山大学公共卫生学院医学统计与流行病学系
出 处: 《数理统计与管理》 2005年第6期62-66,共5页
摘 要: 本文针对多电导水平离子通道的由多个正态分布加权组成的混合分布特点,用EM迭代算法对混合分布中的参数进行极大似然估计,并在此基础上,利用混合分布中最可能的成分判断通道状态,从而还原通道潜在信号,克服了离子通道分析软件PCLAMP中参数估计与状态还原的缺陷。 The current amplitudes of multi -level ion -channel show a mixture distribution which probability density contains m normal density components. In this paper, the maximum likelihood estimates (MLE) were given for the parameters in the mixture distribution via EM algorithm. Based on the mixture distribution, the most possible compo- nent in the mixture distribution is used to classify the channel's states and then to restore the underlying ion -channel signal. These methods in the paper overcome the shortcomings on parameter estimations and classification for states in the software PCLAMP.
领 域: [生物学]