Maximum Likelihood Estimation with Incomplete Multinomial Data.
Themis optimization research program,
TEXAS A AND M UNIV COLLEGE STATION INST OF STATISTICS
Pagination or Media Count:
When sampling from a multinomial population, it may happen either by chance or by design that some of the observations are only partially classified. In the paper the amount of information in the partially classified observation is evaluated, and maximum likelihood estimators are obtained using both completely classified and partially classified data. In general an iterative solution is required, but special cases are identified in which the solution is obtained in closed form. Author
- Statistics and Probability