is yardstick from tidymodels good statistical tool to evaluate metrics like RO, AUC and RMSE

Question

I want to evaluate metric based on interface modeling similar to caret. I used caret for XGBoost, randomFOrest and some time Keras for statistical metric.So I never used tidymodels for metrics evaluation. I want to find the same metric I wrote in title with tidymodels too

I tried with caret mlr3 also and now want to try with tidymodels to check different accuracy metrics. Is it worth checking evaluation metric based on it ? Also I'm expecting to achieve cross validation technique, model comparison, and metric computations ( F1 score, ROC AUC, RMSE, MAE)

Answer 1

You can find a list of metrics, grouped by type, on the yardstick reference page. There is pretty good parity with caret (not sure about mlr3 though).

There is some good reading about using them in chapter 9 of the tidymodels book as well as on tidymodels.org and the workshop notes (such as these slides).