INFORMATION AMOUNT AS A MEASURE CORRECTING ONE'S PRIOR PROBABILITY.
Abstract
Several kinds of definition of information amount or information value have been proposed by several authors. The author attempts to define information amount as a measure of correcting one's prior probability distribution on the state space. By combining two informations concerning the same decision partition, it is shown that the combined information has a greater information amount than the separate ones. When two different types of informations - in the sense they are concerned with different decision sub-partitions of the ultimate decision partition - it is shown how they should be aggregated to obtain a more informative information concerning the ultimate decision partition. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1965
- Accession Number
- AD0617326
Entities
People
- Koichi Miyasawa
Organizations
- University of California, Los Angeles