A method for density estimation based on expectation identities

Joaquín Peralta, Claudia Loyola, Humberto Loguercio, Sergio Davis

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present a simple and direct method for non-parametric estimation of a one-dimensional probability density, based on the application of the recent conjugate variables theorem. The method expands the logarithm of the probability density ln P(x|I) in terms of a complete basis and numerically solves for the coefficients of the expansion using a linear system of equations. No Monte Carlo sampling is needed. We present preliminary results that show the practical usefulness of the method for modeling statistical data.

Original languageEnglish
Title of host publicationBayesian Inference and Maximum Entropy Methods in Science and Engineering
Subtitle of host publicationProceedings of the 36th International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, MaxEnt 2016
PublisherAmerican Institute of Physics Inc.
Volume1853
ISBN (Electronic)9780735415270
DOIs
Publication statusPublished - 9 Jun 2017
Event36th International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, MaxEnt 2016 - Ghent, Belgium
Duration: 10 Jul 201615 Jul 2016

Conference

Conference36th International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, MaxEnt 2016
CountryBelgium
CityGhent
Period10/07/1615/07/16

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Fingerprint Dive into the research topics of 'A method for density estimation based on expectation identities'. Together they form a unique fingerprint.

  • Cite this

    Peralta, J., Loyola, C., Loguercio, H., & Davis, S. (2017). A method for density estimation based on expectation identities. In Bayesian Inference and Maximum Entropy Methods in Science and Engineering: Proceedings of the 36th International Workshop on Bayesian Inference and Maximum Entropy Methods in Science and Engineering, MaxEnt 2016 (Vol. 1853). [110001] American Institute of Physics Inc.. https://doi.org/10.1063/1.4985376