Editing An Information Theory for Preferences

Title: An Information Theory for Preferences

Research Question: How can information theory be applied to preferences to better understand decision-making processes?

Methodology: The study uses an analogy between cumulative probability distributions and normalize utility functions to define a utility density function. This function is non-negative and integrates to 1, forming the basis for a correspondence between utility and probability. The authors introduce the maximum entropy principle, which assigns maximum entropy utility values based on partial preference information.

Results: The research shows that the maximum entropy principle interprets many common utility functions and helps assign utility values based on the information needed for their assignment. The study also introduces joint utility density functions, utility independence, and the concept of mutual preference.

Implications: The application of information theory to preferences has significant implications for decision-making processes. It provides a better understanding of how individuals evaluate different options and make choices under uncertainty. The study also contributes to the development of new methods for preference elicitation and decision-making support systems.

Link to Article: https://arxiv.org/abs/0310045v1
Authors: 
arXiv ID: 0310045v1

[[Category:Computer Science]]
[[Category:Utility]]
[[Category:Information]]
[[Category:Theory]]
[[Category:Preferences]]
[[Category:Decision]]