Ott J
Ann Hum Genet. 1977 May;40(4):443-54.
The likelihood of human pedigree data can be written in such a form as to allow the computation of derivatives. This is done for various parameters in linkage and segregation analysis. The equations for the maximum likelihood estimates are represented in a particularly appealing form which allows iterative solutions. This process is an extension to pedigress of Smith's (1957) counting methods. All these procedures belong to a general class of MLE methods for incomplete data called EM algorithms (Dempster et al. 1976).
人类系谱数据的似然性可以写成一种允许计算导数的形式。在连锁和分离分析中，针对各种参数都进行了这样的处理。最大似然估计的方程以一种特别吸引人的形式表示，这种形式允许迭代求解。这个过程是对史密斯（1957年）计数方法在系谱方面的扩展。所有这些程序都属于一类用于不完全数据的最大似然估计方法，称为期望最大化（EM）算法（邓普斯特等人，1976年）。
