Using the bootstrap in a multivariadte data problem: An example

Technical Report · Tue Aug 01 04:00:00 EDT 1995

DOI:https://doi.org/10.2172/115638· OSTI ID:115638

Glosup, J G; Axelrod, M C

The use of the bootstrap in the multivariate version of the paired t-test is considered and demonstrated through an example. The problem of interest involves comparing two different techniques for measuring the chemical constituents of an sample item. The bootstrap is used to form an empirical significance level for Hotelling`s one-sample T-squared statistic. The bootstrap was selected to determine empirical significance levels because the implicit assumption of multivariate normality in the classic Hotelling`s one-sample test night not hold. The results of both the classic and bootstrap test are presented and contrasted.

• Research Organization: Lawrence Livermore National Lab., CA (United States)
• Sponsoring Organization: USDOE, Washington, DC (United States)
• DOE Contract Number: W-7405-ENG-48
• OSTI ID: 115638
• Report Number(s): UCRL-JC--121819; CONF-9508165--3; ON: DE96001320
• Country of Publication: United States
• Language: English

Related Subjects

40 CHEMISTRY
99 GENERAL AND MISCELLANEOUS
ALGORITHMS
EXPERIMENTAL DATA
ISOTOPE RATIO
MULTIVARIATE ANALYSIS