NLP Modeling with TidyModels and Keras

In this #TidyTuesday video, I go over Animal Crossing reviews and try to predict the review score with the given review text. I explain how to use and perform a Principal Components Analysis (PCA) on text data for a basic exploratory analysis. I also use TF-IDF and Weighted Log-Odds to evaluate if the predictive modeling problem is generalizable. Then, I go over the baseline #NLP model using Lasso regression and then go onto a more complicated shallow black-box model, RandomForest using #TidyModels. Lastly, I go over how to develop a simple neural network model and the intuition of word embeddings such as Word2Vec and BERT.

[aristhotle]

Just found your channel. Very useful videos and on techniques where there's not many (if any) guides available online. Thanks a bunch!

[KyPapie]

Awesome walkthrough! Thanks for recording these, you do an amazing job of explaining code and basic concepts! I have just recently started to get into tidymodels, this has truly inspired me! :)

[sithius]

great video! I don't see a whole lot of R users using neural nets so this was really cool to watch. Thanks.

[rdnyansagar]

Your tutorials are awesome. But I kindly ask that you increase the font size so that it is readable in a normal youtube window. Your font is so small that difficult to read in full-screen mode.