Information-Theoretic Non-Parametric Unimodal Density Estimation.
Abstract
This document presents an information-theoretic method for nonparametric density estimation which guarantees that the resulting density is unimodal. The method inputs data in the form of moment or quantile information and consequently can handle both data derived and non-data derived information. In the non-data derived situation it yields a method for obtaining unimodal Bayesian prior distribution. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1984
- Accession Number
- ADA144522
Entities
People
- Abraham Charnes
- K. Paick
- P. Brockett
Organizations
- University of Texas at Austin