Maximum Entropy Interval Aggregations
Information Theory
2018-05-16 v1 Data Structures and Algorithms
math.IT
Abstract
Given a probability distribution and an integer , we say that is a contiguous -aggregation of if there exist indices such that for each it holds that In this paper, we consider the problem of efficiently finding the contiguous -aggregation of maximum entropy. We design a dynamic programming algorithm that solves the problem exactly, and two more time-efficient greedy algorithms that provide slightly sub-optimal solutions. We also discuss a few scenarios where our problem matters.
Keywords
Cite
@article{arxiv.1805.05375,
title = {Maximum Entropy Interval Aggregations},
author = {Ferdinando Cicalese and Ugo Vaccaro},
journal= {arXiv preprint arXiv:1805.05375},
year = {2018}
}
Comments
To be presented at ISIT 2018