The least-squares combination of correlated data is alive and well

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


The authors discuss the least-squares procedure for averaging any number of correlated data. In fact, they consider the procedure in its general formulation which serves to determine m parameters by n correlated experimental data on different functions of the parameters. If m>n, the same formulation applies to the adjustment of a given library of m correlated parameters by a set of n relevant correlated data which may even be correlated to the given parameters. There are, however, certain restrictions on the application of the least-squares procedure. Combining averages over two partially overlapping sets of data, for instance, yields a result which is not equal to the average over the union of the two sets. The authors offer a general comment on the applicability of the least-squares method.
Original languageEnglish
Title of host publicationInternational Symposium on Nuclear Data Evaluation Methodology
Subtitle of host publicationBrookhaven National Laboratory, Upton, NY, USA, 12-16 October 1992
EditorsCharles L. Dunford
Place of PublicationSingapore
PublisherWorld Scientific Publishing Co. Pte Ltd
Pages231 - 6
Number of pages6
ISBN (Print)981-02-1285-2
StatePublished - 1993

Bibliographical note

correlated data;least-squares;averaging;


  • data analysis
  • least squares approximations
  • radiation detection and measurement


Dive into the research topics of 'The least-squares combination of correlated data is alive and well'. Together they form a unique fingerprint.

Cite this