For a sequence of these functions “t sub n,” define “s sub n” as equal to “t sub n” whenever “t sub n” is greater than “n to the negative one-fourth power” and 0 elsewhere; Joseph Hodges showed that the sequence of these functions “s sub n” has better asymptotic performance than “t sub n,” but is highly irregular. The influence function obtained from Huber’s psi function is used to construct a class of these functions denoted by the letter capital M. When these functions have a high breakdown point, they are robust. These functions’ variance is greater than or equal to one over the Fisher information. These functions either output points or intervals. A common method of obtaining these functions finds the parameters that maximize the log-likelihood. The bias of these functions describes how far off their expected value is. For 10 points, name these functions that approximate the value of a population parameter. ■END■
ANSWER: estimators [accept any kind of estimators; accept statistical estimation, estimates, or maximum likelihood estimation; prompt on statistics] (The sentence about bounding estimator variance references the Cramér–Rao bound.)
<Other Science>
= Average correct buzz position