×

Methodology and algorithms of empirical likelihood. (English) Zbl 0716.62003

Summary: We describe the main features of empirical likelihood and discuss recent developments, including Bartlett correction and location adjustment. Algorithms are provided for implementing empirical likelihood in important cases, for example to means, variances and correlation coefficients. It is shown that empirical likelihood is a serious competitor with contemporary methods such as the bootstrap. Indeed empirical likelihood has several advantages, in that it does not impose prior constraints on region shape, does not require construction of a pivotal statistic, and admits a Bartlett correction which allows very low coverage error. Empirical likelihood deserves a prominent place in the modern statistician’s armoury of computer-intensive tools.

MSC:

62A01 Foundations and philosophical topics in statistics
PDFBibTeX XMLCite
Full Text: DOI