We address the problem of detecting a change that is not persistent but rather starts and terminates at unknown points in time. We revisit the motivation behind several popular maximal likelihood ratio-based rules and investigate their operating characteristics.

翻译：我们处理的问题是如何发现一种变化,这种变化不是持久不变的,而是在未知的时间点开始和结束的。 我们重新审视若干以最大可能性为基础的普遍规则背后的动机,并调查其运作特点。