Conditional event algebra: a new approach
I. R. Goodman
Abstract
I. R. Goodman
Abstract
Recently, algebras of conditional expressions or conditional events compatible with all possible conditional probability evaluations have begun to be developed. Following a brief overview of the main results, two new properties are also demonstration extension of the standard normal disjunctive form expansion to a conditional event setting and a basic connection between the two leading conditional event algebras, showing that one can be considered a weighted averaging of the other in an algebraic sense. However, because of the inherent non-Boolean structure of all of the current approaches, a number of difficulties have arisen, including the issues of higher order conditioning modeling, defining of conditional random variables, and relations to likelihood forms via standard numerically-oriented probability. By taking a fresh approach, based on a product probability space formation, a sounder theory can be established, though it entails significantly more computations.>
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Recently, algebras of conditional expressions or conditional events compatible with all possible conditional probability evaluations have begun to be developed. Following a brief overview of the main results, two new properties are also demonstration extension of the standard normal disjunctive form expansion to a conditional event setting and a basic connection between the two leading conditional event algebras, showing that one can be considered a weighted averaging of the other in an algebraic sense. However, because of the inherent non-Boolean structure of all of the current approaches, a number of difficulties have arisen, including the issues of higher order conditioning modeling, defining of conditional random variables, and relations to likelihood forms via standard numerically-oriented probability. By taking a fresh approach, based on a product probability space formation, a sounder theory can be established, though it entails significantly more computations.>
Key concepts: Event (particle physics), Conditional probability, Regular conditional probability, Extension (predicate logic), Conditional expectation, Connection (principal bundle), Boolean algebra, Computer science