Title: ORSS and Applications
to Inference
Speaker: Prof. N. Balakrishnan
Dept. of Mathematics and Statistics
McMaster University
Hamitlon, Ontario, Canada
Abstract:
In this talk, I will introduce the notion of ORSS (Ordered
Ranked Set
Sample) and first discuss its distributional properties. I
will then
develop Best Linear Unbiased Estimators based on ORSS and
show that
they are considerably more efficient than those based on RSS
for cases
such as normal, logistic, exponential, and generalized geometric.
Next, I will discuss the construction of exact nonparametric
confidence intervals for quantiles and tolerance intervals
based on
ORSS. Finally, I will discuss the MLEs of location and scale
parameters based on ORSS and make a comparison of ORSS and
RSS through
Fisher Information as well as Tukey's Linear Sensitivity Measure.
|
