This package performs Bayesian nonparametric density estimation for exact and censored data via a normalized random measure mixture model. The package allows the user to specify the mixture kernel, the mixing normalized measure and the choice of performing fully nonparametric mixtures on locations and scales, or semiparametric mixtures on locations only with common scale parameter. Options for the kernels are: two kernels with support in the real line (gaussian and double exponential), two more kernels in the positive line (gamma and lognormal) and one with bounded support (beta). The options for the normalized random measures are members of the class of normalized generalized gamma, which include the Dirichlet process, the normalized inversed gaussian process and the normalized stable process. The type of censored data handled by the package is right, left and interval.
|License:||GPL version 2 or later|
The package includes four main functions: MixNRMI1, MixNRMI2, MixNRMI1cens and MixNRMI2cens which implement semiparametric and fully nonparametric mixtures for exact data, and semiparametric and fully nonparametric mixtures for censored data respectively. Additionally, the package includes several other functions required for sampling from conditional distributions in the MCMC implementation. These functions are intended for internal use only.
Barrios, E., Lijoi, A., Nieto-Barajas, L. E. and Prüenster, I.; Contributor: Guillaume Kon Kam King.; Maintainer: Ernesto Barrios <ebarrios at itam.mx>
Barrios, E., Lijoi, A., Nieto-Barajas, L. E. and Prüenster, I. (2013). Modeling with Normalized Random Measure Mixture Models. Statistical Science. Vol. 28, No. 3, 313-334.
Kon Kam King, G., Arbel, J. and Prüenster, I. (2016). Species Sensitivity Distribution revisited: a Bayesian nonparametric approach. In preparation.
1 2 3 4