Variable importance

[Jordan Engelbrecht]

Hello!

I am trying to figure out how to generate a table of environmental variable importance once I've run my model. I've tried running the model code in R and including the line: "var.importance(model)" but I've had no success thus far. I've also tried using the package "maxent" and including the following lines: 

occ_matrix <- cbind(occs_Pp, rep(1, nrow(occs_Pp)))  

maxnet_model <- maxnet(
  p = occ_matrix[,3],  # Presences
  data = as.data.frame(envs_Pp),
  f = maxnet.formula(occ_matrix[,3], envs_Pp)
)

importance <- maxnet::importance(model_Pp)
importance_df <- as.data.frame(importance)
importance_df$Variable <- rownames(importance_df)

This did not work either...

Has anyone perhaps figured out how to do this? 

Your help would be much appreciated!

Thank you

Jordy

[Wallace Ecological Modeling App]

I do not believe the `maxnet` package runs permutation importance, but `ENMeval` might. 

I'll tag Jamie Kass, but would `enm.variable.importance()` or `eval.variable.importance()` work for this?

Bethany

The Wallace Team

[Jordan Engelbrecht]

Thank you so much. enm.variable.importance() isn't working and neither is eval.variable.importance(). I saw on Stack Overflow that someone had the same issue and resolved it using maxent.jar. I tried to change the model algorithm from "maxent" to "maxent.jar" but keep getting errors (see screenshot below). Do you or @Jamie Kass have any advice?

Your help would be greatly appreciated!

[Jordan Engelbrecht]

Thank you so much! Neither enm.variable.importance() nor eval.variable.importance() will work. I saw somewhere that someone managed to generate the variable importance results by running the model with the algorithm "maxent.jar" and not "maxent". I did so, but it still won't work...see code in screenshot attached. Do you or @Jamie Kass have any suggestions? (or anyone for that matter). I have installed Java and the Java execution file maxent. 

Your help would be greatly appreciated!!

On Monday, 05 August 2024 at 17:53:28 UTC+2 Wallace Ecological Modeling App wrote:

[Wallace Ecological Modeling App]

Looks like there's a typo (one I myself make frequently). The two options for setting the algMaxent are maxnet or maxent.jar. Try switching the N and the E.

-Bethany

[Jordan Engelbrecht]

Wow haha. Always helps to have another set of eyes look at your code. Thank you so so much!

--
You received this message because you are subscribed to a topic in the Google Groups "Wallace Ecological Modeling App" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/wallaceEcoMod/9p0BJ4YSD-g/unsubscribe.
To unsubscribe from this group and all its topics, send an email to wallaceEcoMo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/wallaceEcoMod/67637982-3637-4f7d-ab21-0eaf2f6565f3n%40googlegroups.com.

[Mariana Mayte Hérnandez Flores]

Hello! 

Were you able to solve your problem? Did you finally get the contribution of the variables for your models? 

I want to have this information but I have no idea how to do it, could you help me?