Statistical Inference for a Two-State Markov Model of a Single Ion Channel, Incorporating Time Interval Omission
研究了单离子通道两状态马尔可夫模型中，当短暂开放和关闭状态无法被检测时，如何估计平均开放和关闭时间，并比较了矩估计与近似模型下的置信区域。
SUMMARY A two-state Markov model of a single ion channel, in which successive open and closed sojourns follow negative exponential distributions with respective means μO and μC, is considered. Estimation of (μO μC) when brief sojourns in the open and closed states fail to be detected is studied. The moment estimating equations for (μO, μC) have no, one or two solutions, with two being the most common situation. The asymptotic behaviour of method-of-moments estimators of (μO, μC), as the number of sojourns becomes large, is determined, and the resulting confidence regions for (μO, μC) are compared with those derived by Clarke, Milne and Yeo under an approximate model. A new method of discriminating between the two solutions of the moment estimating equations is presented. The theory is illustrated by numerical examples taken from ion channel literature.