Go top
Conference paper information

A parzen-based distance between probability measures as an alternative of summary statistics in approximate bayesian computation

C.D. Zuluaga-Ríos, E.A. Valencia, M.A. Álvarez-López, A.A. Orozco-Gutiérrez

18th International Conference on Image Analysis and Processing - ICIAP 2015, Genoa (Italy). 07-11 September 2015


Summary:

Approximate Bayesian Computation (ABC) are likelihood-free Monte Carlo methods. ABC methods use a comparison between simulated data, using different parameters drawn from a prior distribution, and observed data. This comparison process is based on computing a distance between the summary statistics from the simulated data and the observed data. For complex models, it is usually difficult to define a methodology for choosing or constructing the summary statistics. Recently, a nonparametric ABC has been proposed, that uses a dissimilarity measure between discrete distributions based on empirical kernel embeddings as an alternative for summary statistics. The nonparametric ABC outperforms other methods including ABC, kernel ABC or synthetic likelihood ABC. However, it assumes that the probability distributions are discrete, and it is not robust when dealing with few observations. In this paper, we propose to apply kernel embeddings using a sufficiently smooth density estimator or Parzen estimator for comparing the empirical data distributions, and computing the ABC posterior. Synthetic data and real data were used to test the Bayesian inference of our method. We compare our method with respect to state-of-the-art methods, and demonstrate that our method is a robust estimator of the posterior distribution in terms of the number of observations.


Keywords: Probability Measure; Posterior Distribution; Reproduce Kernel Hilbert Space; Robust Estimator; Approximate Bayesian Computation


DOI: DOI icon https://doi.org/10.1007/978-3-319-23231-7_5

Published in Image Analysis and Processing — ICIAP 2015, vol: Part I, pp: 50-61, ISBN: 978-3-319-23230-0

Publication date: 2015-09-02.



Citation:
C.D. Zuluaga-Ríos, E.A. Valencia, M.A. Álvarez-López, A.A. Orozco-Gutiérrez, A parzen-based distance between probability measures as an alternative of summary statistics in approximate bayesian computation, 18th International Conference on Image Analysis and Processing - ICIAP 2015, Genoa (Italy). 07-11 September 2015. In: Image Analysis and Processing — ICIAP 2015: 18th International Conference, Genoa, Italy, September 7-11, 2015. Proceedings., vol. Part I, ISBN: 978-3-319-23230-0

Request Request the document to be emailed to you.