Compares two models by evaluating their Bayes factor
Usage
Bayes.factor(model1, model2, inter=TRUE)
Arguments
model1
object of the class model "Bayesthresh"
model2
object of the class model "Bayesthresh"
inter
If TRUE, print to scale for interpretation of the Bayes factor
Details
At each step during the Markov chains, the marginal likelihood for a model is evaluated,
conditioning on actual values for the parameters in that step. Bayes factor is then estimated
by the ratios of the arithmetic means of marginal likelihoods from both models. Details of the
implementation can be found in Sorensen and Gianola (2004). For a discussion of the possible
interpretation of Bayes factors, see Jeffreys(1961)
References
SORENSEN, D.; GIANOLA, D. Likelihood, bayesian and MCMC methods in
quantitative genetics. United States of America: Springer, 2004. 740 p.
JEFFREYS, H. Theory of probability. Oxford: Claredon Press, 1961. 470 p.
Examples
data(sensory)
Consumer <- factor(sensory$consumer)
Sacarose <- factor(sensory$sacarose)
# Not run
#### Model 1
# Model with Gaussian link
dex1 <- Bayesthresh(flavor ~ (1|Consumer) + Sacarose, burn = 0, jump = 1,
ef.iter = 10, data=sensory)
summary(dex1)
#### Model 2
# Model with t-Student link
dex2 <- Bayesthresh(flavor ~ (1|Consumer) + Sacarose, burn = 0, jump = 1,
ef.iter = 10, algor=list(algorithm="NC", link="t"),data=sensory)
summary(dex2)
Bayes.factor(dex1,dex2)
Results
R version 3.3.1 (2016-06-21) -- "Bug in Your Hair"
Copyright (C) 2016 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)
R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.
R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.
Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.
> library(Bayesthresh)
Loading required package: lme4
Loading required package: Matrix
Loading required package: MASS
Loading required package: VGAM
Loading required package: stats4
Loading required package: splines
Loading required package: mvtnorm
Loading required package: matrixcalc
Loading required package: coda
Attaching package: 'coda'
The following object is masked from 'package:VGAM':
nvar
> png(filename="/home/ddbj/snapshot/RGM3/R_CC/result/Bayesthresh/Bayes.factor.Rd_%03d_medium.png", width=480, height=480)
> ### Name: Bayes.factor
> ### Title: Bayes factor of the two models
> ### Aliases: Bayes.factor
>
> ### ** Examples
>
> data(sensory)
>
> Consumer <- factor(sensory$consumer)
> Sacarose <- factor(sensory$sacarose)
>
> # Not run
>
> #### Model 1
>
> # Model with Gaussian link
>
> dex1 <- Bayesthresh(flavor ~ (1|Consumer) + Sacarose, burn = 0, jump = 1,
+ ef.iter = 10, data=sensory)
> summary(dex1)
Threshold model with algorithm NC and link Gaussian
Formula: flavor ~ (1 | Consumer) + Sacarose
Deviance: 327.5067
Marginal Log-likelihood:
Post. mean Post.std.dev
-163.7533 7.456134
Random effects:
Post.variance Post.std.dev
Consumer 0.1758407 0.05570589
Residuals 0.3958872 0.04558618
Fixed effects:
Estimate Std. Dev
(Intercept) 0.8252476 0.1889901
Sacarose40 -0.1184004 0.1567156
Sacarose50 -0.1526394 0.1234530
Iteraction Control:
Burn = 0 , Jump = 1 , Iteraction = 10
Time elapsed 0.028 seconds
>
> #### Model 2
>
> # Model with t-Student link
>
> dex2 <- Bayesthresh(flavor ~ (1|Consumer) + Sacarose, burn = 0, jump = 1,
+ ef.iter = 10, algor=list(algorithm="NC", link="t"),data=sensory)
> summary(dex2)
Threshold model with algorithm NC and link t
Formula: flavor ~ (1 | Consumer) + Sacarose
Deviance: 329.0357
Marginal Log-likelihood:
Post. mean Post.std.dev
-164.5179 4.749043
Random effects:
Post.variance Post.std.dev
Consumer 0.2022182 0.08191428
Residuals 0.4227307 0.06547137
Fixed effects:
Estimate Std. Dev
(Intercept) 1.0036404 0.3024412
Sacarose40 -0.1578893 0.1212921
Sacarose50 -0.2290554 0.1899861
Iteraction Control:
Burn = 0 , Jump = 1 , Iteraction = 10
Time elapsed 0.035 seconds
>
> Bayes.factor(dex1,dex2)
Bayes factor for comparison two models
Model 1: flavor ~ (1 | Consumer) + Sacarose
Model 2: flavor ~ (1 | Consumer) + Sacarose
Bayes factor
model1/model2 1.004669
Scale for interpretation of the Bayes factor
--------------------------------------------
B_ij Evidence in favor of M_1
--------------------------------------------
<1 negative (favor of M_2)
1 to 3 doubtfull
3 to 10 substantial
10 to 30 strong
30 to 100 very strong
>100 decisive
--------------------------------------------
Jeffreys(1961)
>
>
>
>
>
>
> dev.off()
null device
1
>