ci.lvl not working for ggpredict multinom plot

[melex]

Hi everyone,

I want to create plots with predicted values and confidence intervals for my multinomial logistic models (multinom function, created with the nnet package). I have numerical predictors on the log scale.

I want to create plots with predicted values and confidence intervals for my multinomial logistic models ( multinom function, created with the nnet package). I have numerical predictors on the log scale.

As janix said his post, though ggeffects() should be compatible with multinom, the plot does not display confidence intervals. If I plot the same data with effects(), I do get the CIs.

What I have tried:

I found this very helpful post on RCommunity in which confidence intervals from the effects package were combined with ggeffects to create a plot.

Plotting confidence intervals with ggeffects for multinom

Hello R-people, I have a question regarding the ggeffects package and its use with multinom functions (from nnet package): I am trying to plot marginal effects for a multinomial regression model. Though ggeffects() should be compatible with multinom, the plot does not display confidence intervals. If I plot the same data with effects(), I do get the CIs. I couldn't find any example for the use of ggeffects with multinom, so I'd be grateful for any suggestion that works for numerical and factor predictor variables. (I also tried the MNLpred package and its mnl_pred_ova function as an alternative, but this one doesn't seem to be able to deal with factors...) Here's a minimal example: libra…

However, (and I'm sure this is my error somewhere) when I run this code I get the following error:

Error in FUN(X[[i]], ...) : object 'L' not found

I have also tried MNLpred package and its mnl_pred_ova function as an alternative, but my predictors are in log scale, and it doesn't seem to be able to deal with these - I get the following error:

Error in eval(parse(text = paste0("data$", xvari))) : 
  attempt to apply non-function

Lastly, I tried using the mnlAveEffPlot function from the DAMisc package, which works, but isn't as customisable as I would like.

I am new to using to R and to this community, so my apologies if this question is very basic or is missing something essential. Here is a small example:

library(nnet)
library(plyr)
library(dplyr)
library(ggplot2)
library(tidyr)
library(sjPlot)
library(DAMisc)
library(ggeffects)
library(scales)
library(MNLpred)
library(effects)

df <- data.frame(response = c("1 Better", "1 Better", "1 Better", "2 Medium", "2 Medium", "2 Medium", "3 Worse", "3 Worse", "3 Worse"),
                 count = c(1000, 2000, 4000, 6000, 10000, 3000, 6000, 5000, 11000))
mod1 <- multinom(response ~ log(count),
                         data = df)
summary(mod1)

pred <- ggpredict(mod1, terms = "count") 
plot(pred)

Thanks for your help!