2.1.24.5.2.3 ocmath_para_gaussmix


Description

Estimate parameters for Gaussian Mixture distribution

Syntax

int ocmath_para_gaussmix(int iSize, const double * pData, double * pPara, int * pnG, GaussMixVar nVar, bool bOpt, int riter, double * pAD = NULL, double * pBIC = NULL, int nPercents = 0, const double * pPercents = NULL, double * pXp = NULL, double * pPval = NULL, bool bMC = false)

Parameters

iSize
[input]input data size
pData
[input]input data
pPara
[output]weight, mean, standard deviation for each component in the mixture
pnG
[input/output]number of groups
nVar
[input]
enum GaussMixVar
{
	var_Group      = 0,
	var_Pooled     = 1,
	var_GroupDiag  = 2,
	var_PooledDiag = 3,
	var_Overall    = 4,
};
bOpt
[input]true to find the best number of components, ng: the maximum number of components; false to use the specified number of components ng
riter
[input]membership probabilities are rounded to 0 or 1 after every riter iterations, -1 means not rounded
pAD
[output]Anderson-Darling statistic
pBIC
[output]BIC
nPercents
[input]number of Percents
pPercents
[input] Percents
pXp
[output]XP
pPval
[output]P-value
bMC
[input]true to calculate AD and Pval by Monte Carlo

Return

return 0 on successful exit; if an error or warning has been detected, then it returns the specific error or warning code.

Examples

Remark

See Also

Header to Include

origin.h

Reference