No video :(

Evaluate multiple ML approaches for spam detection

Подписаться 15 тыс.

Просмотров 3,1 тыс.

50% 1

In this screencast, we use tidymodels workflowsets to try out multiple modeling approaches for a #TidyTuesday dataset on spam email. We finish off with how to create a deployable model object and set up an API for our model. Check out the code on my blog: juliasilge.com...

Опубликовано:

28 авг 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 12

@MrAbhimufc Год назад

Been here from when your introduction had stack overflow in it. You were a tutor once and now just part of my weekend analytical routine. Thank you.

@tiernanmartin3729 Год назад

I always learn so much from your tutorial videos! It's exciting to watch the tidymodels framework becoming more and more fleshed out.

@cheninitayeb6590 Год назад

Thank you very much for this screencast ,I hope to see your lecteur twice a month ,again ,many thanks .

@darmaw22 Год назад

Many thanks for yet another great lecture!

@HM-wo6ic Год назад

Great video. Thanks for the clarification on random forests behavior regarding tuning and trees.

@atlantaguitar9689 11 месяцев назад

Great video. Thanks. Students still present questions like "when do I juice? Before or after baking? Why would I juice something that is to be baked? Can I bake something that isn't prepped? If I use workflows it looks like I can avoid the cooking analogies altogether since they don't seem to naturally extend into workflows", etc. Over time I think jumping straight into the workflows helps show students (and teachers) the framework from a top down approach. We can the use functions at a more basic level to examine what's happening underneath the hood.

@JuliaSilge 11 месяцев назад

Yep, that is how we teach tidymodels in our book, workshops, etc. Take a look at _Tidy Modeling with R_. We introduce models, workflows, and recipes, in Chs 6, 7, and 8, and don't talk about `prep()` or `bake()` at all until Ch 16: www.tmwr.org/dimensionality#recipe-functions We don't teach with `juice()` at all, and just reserve that for expert and internal use.

@Jakan-sf3xj 8 месяцев назад

Thank you for the great video. I have one question, assuming the best model was one of the tuned random forest models, how would we extract the parsnip object to see the tuned hyperparameters i.e mtry and min_n?

@JuliaSilge 8 месяцев назад

You might check out the different "extract" functions in tidymodels. You can do `extract_fit_parsnip()` but you can also do `extract_parameter_set_dials()` to get the hyperparameters directly: hardhat.tidymodels.org/reference/hardhat-extract.html

@matthewschorr1038 Год назад

Can you please go over the time_split method for time-series data ?

@hesamseraj Год назад

Thank you very much for such a great screencast. Has the norm changed in using the pipe sign (%>%)? I try to use your style while I code in R. What are the benefits of using |>? Is there a shortcut for it?

@JuliaSilge Год назад

There are two posts I think are helpful for understanding the differences between the base R and magrittr pipes: - ivelasq.rbind.io/blog/understanding-the-r-pipe/ - www.tidyverse.org/blog/2023/04/base-vs-magrittr-pipe/ For most cases, they act the same so no need to stress too much! You can change the shortcut in RStudio to the base R one, if you prefer: stackoverflow.com/questions/68667933/how-to-fast-insert-new-pipe-operator-in-rstudio