Premium
Pairwise Rank‐Based Likelihood for Estimation and Inference on the Mixture Proportion
Author(s) -
Heller Glenn,
Qin Jing
Publication year - 2001
Publication title -
biometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.298
H-Index - 130
eISSN - 1541-0420
pISSN - 0006-341X
DOI - 10.1111/j.0006-341x.2001.00813.x
Subject(s) - statistics , mathematics , pairwise comparison , inference , estimator , rank (graph theory) , population , nonparametric statistics , sufficient statistic , computer science , artificial intelligence , combinatorics , demography , sociology
Summary. We consider the problem of estimation and inference on the mixture parameter in the two‐sample problem when sample data from the two distributions as well as from a third population consisting of a mixture of the two are used. Under a general nonparametric model, where the relationship between the two populations is unspecified, we develop a pairwise rank‐based likelihood. Simultaneous inference on the mixture proportion and a parameter representing the probability an observation from one population is greater than an observation from the other population is based on this likelihood. Under some regularity conditions, it is shown that the maximum pairwise rank likelihood estimator is consistent and has an asymptotic normal distribution. Simulation results indicate that the performance of this statistic is satisfactory. The methodology is demonstrated on a data set in prostate cancer.