1. Likelihood Methods for Combining Tables of Data.
- Author
-
Tolley, H. Dennis and Fellingham, Gilbert W.
- Subjects
- *
MATHEMATICS in insurance , *DEATH rate - Abstract
Often times actuaries, health policy planners and managers must draw conclusions from several related sources of data. Particularly for actuaries, there is often a need to construct a table based on the experience of several blocks of business or on several different tables of data. Although sophisticated methods for using the resulting tables of data have been developed, often that table representing the information combined from several sources is constructed in a relatively ad hoc manner. This paper presents a likelihood method of combining tables of data from several sources. The methods are based on techniques dating at least back to the mid 1940's, but have been made more feasible with advances in computational power. The method presented allows for situations where the data sources have aggregated the data at different levels. In this case questions of estimability are examined using the implicit function theorem. The method is illustrated with an example. [ABSTRACT FROM AUTHOR]
- Published
- 2000
- Full Text
- View/download PDF