Function hnp - Package hnp of R

Question

Function hnp - Package hnp of R

Navigation

#1 by (0 votes)

2

My template was created with the glmer function, of pacete lme4 . It is a generalized linear mixed model with Binomial response.

form=status~(A1+A2+  B1 + ... + B10 +  (1 | escola))

Mod1 <- glmer(form, data = dadosesc, family = binomial, control = glmerControl(optimizer =  "optimx", optCtrl=list(method="nlminb")), 
        nAGQ = 25)

Being status 0 or 1, A1 and A2 and school categorical and other continuous.

I'm trying to use the hnp function, from the hnp package, to create a simulated envelope as follows

hnp(Mod1, xlab="Half-normal scores", ylab="Deviance residuals", pch=3)

but is displaying the following error:

Error in simulate(object)[, 1][, 1] : incorrect number of dimensions.

Only when using a simple binomial formed with GLM.

r glmer

asked by anonymous 05.12.2016 / 20:04

1 answer

Generate multiple samples Changing MySQL with ajax and PHP

score 0 · Answer 1

No hnp help (pages 11 and 12 of this document) , you can see the list of templates that can be used. Templates made with the glmer package are not listed here.

By help, you can also see that it is possible to construct this graph for an unimplemented model. In this case you will need to provide three functions:

simfun: simulates random numbers with model distribution
fitfun: adjust a template again
diagfun: calculates the diagnostic measure to which you want to make the envelope.

For example, consider the following database:

df <- data_frame(
  x1 = runif(100, min = 0, max = 0.5),
  x2 = runif(100, min = 0, max = 0.5),
  y = as.factor(rbinom(n = 100, size = 1, prob = x1 + x2))
)
df
# A tibble: 100 × 3
           x1         x2      y
        <dbl>      <dbl> <fctr>
1  0.05263752 0.44304477      1
2  0.23866105 0.38094325      1
3  0.16474934 0.12541530      0
4  0.47944698 0.46145428      1
5  0.31256457 0.01484219      1
6  0.28411475 0.17338685      1
7  0.34083505 0.35194770      0
8  0.30603019 0.40298573      1
9  0.22193441 0.44697690      1
10 0.36469374 0.34312517      1
# ... with 90 more rows

In this example y is directly related to x1 and x2 .

By setting a mixed template in this case, and running the function hnp I get the same error as you:

fit_glmer <- glmer(y ~ x1 + x2 + (1 | x1), data = df, family = "binomial")
hnp(fit_glmer)
Binomial-normal model 
Error in '[.default'(simulate(object)[, 1], , 1) : 
  incorrect number of dimensions

This happens because no method is implemented to deal with objects of type glmer . Let's then implement the functions needed to construct the graph:

diagfun <- function(obj) {
  resid(obj, type="pearson")
}

simfun <- function(n, obj) {
  as.factor(rbinom(n, size = 1, prob = predict(obj, type = "response")))
}

fitfun <- function(y.) {
  glmer(y. ~ x1 + x2 + (1 | x1), data = df, family = "binomial")
}

You can now call the hnp function this way:

hnp(fit_glmer, newclass=TRUE, diagfun=diagfun, simfun=simfun, fitfun=fitfun)

It will produce the following chart:

Of course, this is just an example; you will need to adapt this code to your specific problem. At first, I think the only place you need to tinker is in the fitfun function. You may need to tinker with the diagfun function if you want the type of residue not to be Pearson for example (I saw in the legend of your graph that you use deviance residues).