Predictability and Information Part I: Measures of Predictability
The connection between predictability and information theory is discussed with the aim of clarifying the various metrics which have been proposed in the literature. This discussion begins with the following definition of predictability: an event is predictable if the posterior distribution, after a set of observations becomes available, differs in any way from the prior distribution which is known before observations become available.
Saved in:
Main Author: | |
---|---|
Format: | Preprint biblioteca |
Language: | English |
Published: |
2003
|
Subjects: | Information retrieval, |
Online Access: | http://hdl.handle.net/1834/470 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | The connection between predictability and information theory is discussed with
the aim of clarifying the various metrics which have been proposed in the literature. This
discussion begins with the following definition of predictability: an event is predictable
if the posterior distribution, after a set of observations becomes available, differs in any
way from the prior distribution which is known before observations become available. |
---|