Login

News

Adjusted BIC

Posted on
Picture of user. trinewa Joined: 11/30/2009
Forums
Behavioral Genetics Models
Is there any way to get the 'adjusted BIC' statistic output in OpenMx?
Replied on Mon, 02/21/2011 - 10:13
Picture of user. Ryne Joined: 07/31/2009

It'll be released in the next binary, along with other added fit indices in 'summary'. For now, you can calculate it by hand with the following formula or function:

sBIC (sample-size adjusted BIC) = -2LL + log((n+2)/24)

or

trinewa <- function(model){
  neg2LL <- model@output$Minus2LogLikelihood
  adjN <- (model@data@numObs+2)/24
  return(neg2LL + log(adjN))
}
Log in or register to post comments
Replied on Thu, 02/24/2011 - 18:06
No user picture. Sabha Joined: 05/18/2010

Along the same lines, is there any option available at the moment to compute pval and chi square values for ACE model fit. I have pasted an example of the summary output as an example.

> univACESummX

free parameters:
name matrix row col Estimate Std.Error
1 a11 ACE.a 1 1 -4.302609e-01 0.1389135
2 c11 ACE.c 1 1 -3.342676e-08 0.9382380
3 e11 ACE.e 1 1 5.466756e-01 0.1022683
4 mean ACE.Mean 1 1 3.361335e-08 0.1264393

confidence intervals:
lbound estimate ubound
univACE.StdA[1,1] 3.815964e-15 3.825054e-01 0.7105036
univACE.StdC[1,1] 9.630561e-17 2.308673e-15 0.6585438
univACE.StdE[1,1] 2.894965e-01 6.174946e-01 1.0000000

observed statistics: 38
estimated parameters: 4
degrees of freedom: 34
-2 log likelihood: 78.70771
saturated -2 log likelihood: NA
number of observations: 19
chi-square: NA
p: NA
AIC (Mx): 10.70771
BIC (Mx): -10.70161
adjusted BIC:
RMSEA: NA
timestamp: 2011-02-24 14:53:16
frontend time: 0.608 secs
backend time: 0.03200006 secs
independent submodels time: 0 secs
wall clock time: 0.6400001 secs
cpu time: 0.6400001 secs
openmx version number: 1.0.4-1540

Log in or register to post comments
  Login

News