o1iv3r / featureimpcluster Goto Github PK

Use foreach in PermMisClassRate and/or FeatureImpCluster.x

Hi, i've found your approach very interesting. If there is any interest, i'd like to help implementing a python version. Are there any papers explaining your approach so i can use as a reference?

Thanks!

This is a really useful tool - thankyou!

I have quite a lot of data, and when I come to plotting it gets a bit overwhelming to visualise...!

I have successfully filtered the "FeatureImp_res " to plot the misClassRate for the features with the highest values.

However, it would be good to subset the "res" before barplotting - i.e. so you can plot one cluster at a time, or again filter based on the values. I will persist, but thought I would drop an issue incase there is a quick answer waiting for me :)

Thanks!

from the code: attempt::stop_if(is.null(predFUN), "Provide prediction function")

according to the help stop_if(.x, .p = isTRUE, msg = NULL)

thus the above should be attempt::stop_if(is.null(predFUN), msg="Provide prediction function")