Measures of Dependence for Evaluating Information in Censored Models.
FLORIDA STATE UNIV TALLAHASSEE DEPT OF STATISTICS
Pagination or Media Count:
Measures of information in censored models are developed by adapting measures of dependence between the lifetime variable and the observed variable. Some common notions of bivariate dependence and coefficients of divergence are used to derive these classes of measures. It is shown that most of the measures of bivariate dependence have the fundamental property that as censoring decreases stochastically, the information increases. An exception occurs when dependence is defined in terms of association. Conditions under which the coefficients of divergence enjoy the fundamental property are established. Author
- Statistics and Probability