Explore subword features in machine learning for text analysis through a practical demonstration using tidymodels to predict Hawaiian post offices based on their names. Learn how to work with the #TidyTuesday dataset, preprocess text data, implement a recipe for feature engineering, tune hyperparameters, and apply a linear SVM model. Gain insights into evaluating model performance through metrics and estimates, all while following along with the code available on Julia Silge's blog.
Overview
Syllabus
Introduction
Dataset
Post office names
What states post office names
Training and testing
Recipe
Hyperparameter
Output
Linear SVM
Results
Metrics
Estimate
Taught by
Julia Silge
