Abstract
The maximum entropy method - used to infer probabilistic models from data - is a special case of Bayes's model inference prescription which, in turn, is grounded in basic propositional logic. By contrast to the maximum entropy method, the compatibility of nonadditive entropy maximization with Bayes's model inference prescription has never been established. Here we demonstrate that nonadditive entropy maximization is incompatible with Bayesian updating and discuss the immediate implications of this finding. We focus our attention on special cases as illustrations.
Original language | English (US) |
---|---|
Article number | 052149 |
Journal | Physical Review E - Statistical, Nonlinear, and Soft Matter Physics |
Volume | 90 |
Issue number | 5 |
DOIs | |
State | Published - Nov 24 2014 |
Externally published | Yes |
ASJC Scopus subject areas
- Statistical and Nonlinear Physics
- Statistics and Probability
- Condensed Matter Physics