Hi I want to know is there any machine learning basic essential framework other than caret?

[yuanqingye]

Hi I think caret is one of the best frame work I am using in R
And I think I use it most.

But sometimes I couldn't find some model get equipped in it,
So I just wondering except for caret, is there any other machine learning frame work I could turn to, if I couldn't find some model I want to use in it ?

I think the author of this famous package may know this best.

Thank you very much.

[nks1974]

Hi yuanqingye, see https://topepo.github.io/caret/using-your-own-model-in-train.html

I would stick to caret, and additionally run models outside caret, and take the model objects apart. You can also see qeML which is a much simpler concept. There are also mlr3 and tidymodels, but they did not beat caret last I looked.

Additionally, I would warn against model and algo fascination. There is no holy grail. Just a bunch of toys.

Key is data load efficiency to join many different data sets, cleaning, transformation and other preparation.

[yuanqingye]

Hi yuanqingye, see https://topepo.github.io/caret/using-your-own-model-in-train.html

I would stick to caret, and additionally run models outside caret, and take the model objects apart. You can also see qeML which is a much simpler concept. There are also mlr3 and tidymodels, but they did not beat caret last I looked.

Additionally, I would warn against model and algo fascination. There is no holy grail. Just a bunch of toys.

Key is data load efficiency to join many different data sets, cleaning, transformation and other preparation.

Sure, will take a look.
Thank you for the important info.
Really appreciated.