Explore the evolution of art media in the Tate collection through a 42-minute tutorial on training regularized regression models with text features using tidymodels. Learn to analyze changes in artistic mediums over time, perform model diagnostics, and interpret results. Dive into data preprocessing, token filtering, feature engineering, and variable importance. Discover how to visualize predictions, assess model performance, and gain insights into artistic trends. Access accompanying code on Julia Silge's blog for hands-on practice and further exploration.
Overview
Syllabus
Introduction
Data set overview
The medium column
The artwork column
The distribution over time
Residuals
Biases
Materials
Preprocessing Data
Training Data
Token Filter
Transform to Matrix
Feature Preprocessing
Sparse Data
Change Range
Training
Results
RMSE
Penalty
Variable importance
Arranging by importance
Making a graph
Scales
Collect predictions
Collect predictions on art final
Filter predictions
Conclusion
Taught by
Julia Silge
