s-news
[Top] [All Lists]

mclust question

To: <s-news@lists.biostat.wustl.edu>
Subject: mclust question
From: "Huss, Matthew" <mhuss@creighton.edu>
Date: Thu, 18 Oct 2007 15:17:26 -0500
Thread-index: AcgRw+Z6gmvIcK43Tw+ko6i9HKZ8ug==
Thread-topic: mclust question
I am attempting to perform a model-based cluster analysis using Mclust and Splus on a dataset of 7 variables and 111 subjects. I am an getting warning message and I was wondering if anyone had any suggestions? 
 
Matt
 
 
Matthew T. Huss, Ph.D.
Associate Professor
Creighton University
Department of Psychology
2500 California Plaza
Omaha, NE 68178
(402) 280-3773
(402) 280-4748 (fax)
 
 

The warning message is as follows:

Warning messages:

1: reciprocal condition estimate falls below threshold in: meVVV(structure(c(-1.15031, 0.21653, 2.08961, 1.22625, 1.51342, -1.00948, ....

2: reciprocal condition estimate falls below threshold in: meVVV(structure(c(-1.15031,0.21653, 2.08961, 1.22625, 1.51342, -1.00948, ....

3: reciprocal condition estimate falls below threshold in: mstepVVV(data, z,eps = eps, equal = equal)below and I attached the syntax for some context.

 

The error message occurs after the following line of the full syntax below:

> MDQBIC <-emclust(MDQ)

 

 

 

FULL SYNTAX

> library (mclust)

> c1 <- Guck[[1]]

> c2 <- Guck[[2]]

> c3 <- Guck[[3]]

> c4 <- Guck[[4]]

> c5 <- Guck[[5]]

> c6 <- Guck[[6]]

> c7 <- Guck[[7]]

> MDQ <- cbind (c1, c2, c3, c4, c5, c6, c7) MDQBIC <- emclust (MDQ)

>MDQBIC MDQmclust <- mclust (MDQ) MDQBIC MDQsum <-summary (MDQBIC, MDQ)

>MDQsum

>MDQEEE3 <-summary (MDQBIC, MDQ ,G = 3, modelNames = "EEE")

>MDQEEE3

>MDQEEE3means <-(MDQEEE3$mu)

>MDQEEE3means

>MDQEEE3group <-(MDQEEE3$classification) MDQEEE3group

 
<Prev in Thread] Current Thread [Next in Thread>
  • mclust question, Huss, Matthew <=