scientific article; zbMATH DE number 1049276

Authors William W. S. Chen, Narayanaswamy Balakrishnan

Publication date 18 August 1997

best linear unbiased estimators failure times inverse Gaussian distributions order statistics reliability

General reference works (handbooks, dictionaries, bibliographies, etc.) pertaining to statistics (62-00) Point estimation (62F10) Order statistics; empirical distribution functions (62G30) Reliability and life testing (62N05) Statistical tables (62Q05) Probabilistic methods, stochastic differential equations (65C99)

Review text

The class of inverse Gaussian distributions is currently being used quite commonly as a life-time model in reliability studies. Chapter 2 of the book gives different forms of inverse Gaussian distributions and some basic properties. Chapter 3 gives an introduction to order statistics and describes computational procedures used for the determination of the means, variances and covariances. Values of means and variances of order statistics for samples sizes up to 25, for the value of the shape parameter being 0.0 (0.1) 2.5, computed by a numerical (single) integration procedure, are presented in Tables 1 and 2, respectively, and those of covariances for the above values of the shape parameter, computed by a numerical (double) procedure, are presented in Table 3. Chapter 4 gives the necessary formulas for the best linear unbiased estimators of the location and scale parameters (assuming the shape parameters to be fixed) based on complete as well as Type-II censored samples. The coefficients in the best linear unbiased estimators for sample sizes up to 25, for the values of the shape and scale parameters being 0.0(0.1)2.5, for various levels of censoring, are presented in Tables 4 and 5, respectively. Table 6 gives the values of the variances and covariances of these best linear unbiased estimators. Chapter 5 presents some numerical examples in order to illustrate a direct application of these tables. The best linear unbiased estimators so derived can also be used successfully to develop best linear unbiased predictions of the failure times corresponding to the censored units. This is demonstrated in Chapter 6 and is also illustrated with some numerical examples. Another very useful application of means of order statistics is in constructing Q-Q plots and in developing formal correlation-type goodness-of-fit tests in order to verify the validity of inverse Gaussian distributions for the observed data. Chapter 7 illustrates this use of means of order statistics presented in Table 1 through numerical examples considered earlier. The book will be very useful to research workers interested in inverse Gaussian distributions.

Recommendations

Cited in

(6)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4348481)