EM 算法是 Dempster，Laind，Rubin 于 1977 年提出的求参数极大似然估计的一种方法，它可以从非完整数据集中对参数进行 MLE 估计，是一种非常简单实用的学习算法。这种方法可以广泛地应用于处理缺损数据，截尾数据，带有讨厌数据等所谓的不完全数据(incomplete data)。需要weka的算法包支持。-EM algorithm is Dempster, Laind, Rubin in 1977 for the parameters proposed by maximum likelihood estimation of a method, it can focus from non-complete data MLE of the parameters estimated, is a very simple and practical learning algorithm. This method can be widely applied to deal with defect data, censored data, with data such as the so-called hate incomplete data (incomplete data).