majority population

[Dovedale]

Пожалуйста, кто знает, помогите перевести majority population

The bias, efficiency, and robustness properties for selected location estimators are deduced for 24 scenarios: four majority populations of ‘‘good’’ measurements contaminated with six different populations of ‘‘bad’’ measurements.

This study evaluates the behavior of the above location estimators applied to sets of ‘‘measurements’’, each set drawn mostly from one of several ‘‘majority’’ populations but mixed with a modest proportion of draws from one of several ‘‘contaminant’’ populations.

Заранее спасибо.

[Dovedale]

Статистики говорят, что это большая часть совокупности, но ведь тогда было бы population majority????

Пожалуйста, уважаемые переводчики, выскажите свое авторитетное мнение.

[Stingray_FM]

навскидку - "основные / преобладающие популяции". контексту подкиньте - биология???

[coucoushkina]

Очень осторожный вариант
"титульное население"

[Dovedale]

Спасибо всем, кто откликнулся!

Статья из области ключевых сличений (метрология). Речь идет о том, как лучше оценивать межлабораторные данные, содержащие аномальные значения неопределенности.

Статья называется: A comparison of location estimators for interlaboratory data contaminated with value and uncertainty outliers

Among the interesting performance properties of location estimators are the average bias (expected difference of the sample-based location estimate, ^m; from the true location of the population, m), maximum bias (worst-case
difference, for a given definition of ‘‘worst’’), efficiency (a measure of the expected variance of the ^m for repeated samplings of the population), and robustness (sensitivity of the estimate to violations of the assumptions). While
there are a number of ways to quantitatively evaluate these properties, this report uses a form of computationally intensive ‘‘Monte Carlo’’ simulation using a reasonably large number of sets of ‘‘measurements’’ generated from
binary mixtures of well-defined distributions. The bias, efficiency, and robustness properties for selected location estimators are deduced for 24 scenarios: four majority populations of ‘‘good’’ measurements contaminated with six different populations of ‘‘bad’’ measurements.

There are ‘‘multi-pass’’ approaches to location estimation that first identify inconsistent or ‘‘outlier’’ data and then use an estimator matched to the structure of the putatively consistent remaining data [3, 11]. Given that
explicit exclusion of technically valid data can be fraught with complications (pragmatic as well as technical), the intrinsic performance characteristics of the estimators themselves are of greatest interest in ‘‘single-pass’’ estimation
where all technically valid data are utilized. However, knowledge of estimator performance characteristics can help define the nature and rigor of the
‘‘consistency’’ required for confident use of a given estimator.

The majority and contaminant populations considered in this report are intended to be at least plausibly appropriate to key comparison data sets. It is hoped the performance of the various estimators examined herein will be increasingly relevant for more typical interlaboratory studies as MU estimation of chemical measurands becomes more routine.