▍1. EM算法(讲解+程序)
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、截尾数据以及带有噪声等所谓的不完全数据，可以具体来说，我们可以利用EM算法来填充样本中的缺失数据、发现隐藏变量的值、估计HMM中的参数、估计有限混合分布中的参数以及可以进行无监督聚类等等。(EM algorithm is a method for maximum likelihood estimation of parameters proposed by Dempster, Laind and Rubin in 1977. It can estimate parameters from incomplete data sets by MLE. It is a very simple and practical learning algorithm. This method can be widely used in dealing with the so-called incomplete data such as defective data, censored data and noisy data. Specifically, we can use EM algorithm to fill missing data in samples, find hidden variable values, estimate parameters in HMM, estimate parameters in finite mixed distribution, and so on. Unsupervised clustering and so on.)
