Editor's special invited paper: On the efficient score vector in sequential monitoring

摘要

The use of he efficient score statistic in sequential monitoring, procedures is reviewed and analysed. Various models that arise in applications are considered. The efficient score vector has the same optimality property as the generalized likelihood ratio, but it has a imple-structure requiring fewer estimations, and this is especially important when the data have a complicated structure. Furthermore, ith the efficient score vector it is possible to detect which component the parameter vector is different from the hypothetical or the historical value. The problems we solve here were the subjects of eparate publications, but with the new methodology they are easily calculated examples.