phenomenon is one of the most striking phenomena in modern statistics.
For about a decade, since 1979, I worked on shrinkage estimation,
especially in relation to Stein's phenomenon from the point of view of
decision theory. Starting from around 1990, my research has focused on
construction of confidence intervals (using empirical Bayes approach at
times) and prediction intervals for various types of problems:
estimation of ratio parameters in regression, measurement error models,
bioequivalence problems, estimation of order restricted parameters and
prediction based on a factor analysis model (HELP). Statistical
intervals, such as confidence intervals, prediction intervals and
tolerance intervals, are of major importance in statistics. By
providing intervals, statisticians demonstrate the possible range of
parameters of interest or future observations.
Courses & Students Advised