Klotz J H
Biometrics. 1978 Jun;34(2):283-7.
A specialization of the biased coin clinical trial randomization schemes of Pocock and Simon (1975) and Efron (1971) is proposed which gives guidelines for the coin biasing probabilities of assignment to treatment. The scheme parameterizes the often conflicting goals of random treatment assignment and allocation for stratum balance in terms of a single trade-off parameter specifying a desired degree of determinism. Treatment radomization probabilities are obtained iteratively in an exponential form which maximizes randomization entropy subject to a specified balance constraint.
本文提出了一种对Pocock和Simon（1975年）以及Efron（1971年）的有偏硬币临床试验随机化方案的特殊化方法，该方法给出了治疗分配硬币偏向概率的指导方针。该方案根据一个指定所需确定性程度的单一权衡参数，对随机治疗分配和分层平衡分配这两个通常相互冲突的目标进行参数化。治疗随机化概率以指数形式迭代获得，在指定的平衡约束下使随机化熵最大化。
