DPGMMclus: DPGMMclus S3 Method

In frimane/SolarClusGnr: Solar irradiance time-series clustering and downscaling

Description

Nonparametric Bayesian Dirichlet-Gaussian clustering of daily clearness index distributions. It can be also used to perform any data clustering of class matrix other than irradiance data of class SIRData.

Usage

DPGMMclus(obj, n.iter, n.burn)

Arguments

obj	object of class SIRData (see SIR_Data) or a matrix object containing the clearness index distributions.
n.iter	numeric(1) represents the number of iterations
n.burn	numeric(1) represents the number of burn-in iterations (ignored iterations)

Value

an object of clusData class containing:

cl_seq	numeric vector represents the class sequence.
like_par	list0object represents the parameters of each class. 1st element is the mean (numeric vector), 5th element is the precision matrix (inverse of the co-variance matrix) of the class.
lik_con_par	numeric(1), represents the inferred concentration parameter alf.
Con_par_sam_seq	numeric vector, represents the posterior distribution of alf.
seq_cl_num	numeric vector, represents the distribution of the class numbers.
bet_seq	numeric vector, represents the destribution of the beta parametrer of the precision matrix.
like_cl	numeric vector, represents the number of elements of each class.
calc_time	numeric(1) represents the computing time consumed.

Author(s)

Azeddine Frimane [e-mail]; [e-mail]

Examples

 
# The example and data below are just to give an idea of how the script works and not to judge the performance of the method.
  
data("SIRData_obj")

newClustering <- DPGMMclus(SIRData_obj, n.iter = 1000, n.burn = 500)
# for class ploting see \code{\link{clPlot}}

frimane/SolarClusGnr documentation built on May 8, 2019, 8:58 p.m.