ESTIMATING AND TESTING WITH THE BHATTACHARYA PLOT

被引:1
作者
BOESE, FG [1 ]
机构
[1] MAX PLANCK INST BIOCHEM,W-8033 MARTINSRIED,GERMANY
关键词
D O I
10.1016/0167-9473(91)90057-9
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
Mixtures of subpopulations of biological particles are appropriately described by finite mixture densities. Often, samples drawn from such populations are condensed in the form of histograms. In the exploratory phase of a parametrical data analysis, one wishes to obtain first estimates for the parameters of such mixtures. A parameter of special significance is the number of subpopulations. For the case of univariate normal mixtures, an explanatory graphical device called Bhattacharya Plot is presented. It allows one to obtain the required first estimates and provides assistance in the test of the normality of the mixture. The degree of overlapping in the histogram must not be too severe.
引用
收藏
页码:111 / 124
页数:14
相关论文
共 13 条
[1]   A SIMPLE METHOD OF RESOLUTION OF A DISTRIBUTION INTO GAUSSIAN COMPONENTS [J].
BHATTACHARYA, CG .
BIOMETRICS, 1967, 23 (01) :115-+
[2]  
Boor CD., 1978, PRACTICAL GUIDE SPLI
[4]  
Everitt B, 2013, FINITE MIXTURE DISTR
[6]   AN ALGORITHM FOR DECOMPOSITION OF A DISTRIBUTION INTO GAUSSIAN COMPONENTS [J].
GREGOR, J .
BIOMETRICS, 1969, 25 (01) :79-&
[7]  
Hald A, 1949, SKAND AKTUARIETIDSKR, V32, P119
[8]   ESTIMATION OF PARMAETERS FOR A MIXTURE OF NORMAL DISTRIBUTIONS [J].
HASSELBLAD, V .
TECHNOMETRICS, 1966, 8 (03) :431-+
[9]  
Hazen A, 1914, T AM SOC CIVIL ENG, V77, P1539, DOI DOI 10.1061/TACEAT.0002563
[10]   COMPUTER-GRAPHICS COMES TO STATISTICS [J].
KOLATA, G .
SCIENCE, 1982, 217 (4563) :919-920