TidyX Episode 79 | Tidymodels - Cross-Validation and Metrics

TidyX Episode 79: Tidymodels - Cross-validation and Metrics

The third episode on tidymodels, we continue our data prep and model training by exploring cross-validation and metric evalidation. Ellis and Patrick show to set up a 5-fold cross validation set on your training split as well as fitting a tidymodels workflow! We finally show how to display and extract model fitting evaluation metrics.

Links:

Open an issue on the TidyX Github page!
TidyX Code:

[Nloon]

Very informative, thank you for this!

[shamsulhoquekhan]

I am having hard time understanding why you did last_fit? where is this last-fitted model coming from? Also, you did not use the result from cross-validation here. how this all connect? I am new so trying to understand all this. Thank you for this video btw. I am following the entire series.

[haraldurkarlsson]

Here is where I get confused. Does it not matter when you impute the NAs? You appear to impute only the training data and not the testing data. Why not impute the whole data set first before the split?

[haraldurkarlsson]

If you are looking for data set you might check out the TidyTuesday data sets - plenty there.

[terraflops]

@4:03, you could have used glimpse(df) or used the {dlookr} pkg

these Tidy Models videos has sparked my interest in reading the books, mostly for understanding recipes and having cute baking variable names.

--- for these models, does the whole dataset need to be numeric? (all columns have integer/ float values)

nice video. I found the explanation of RSME etc helpful as to what it means, maybe next video mutate a column for a grade ?