Overview
Learn how to tune hyperparameters for random forest models using tidymodels in this comprehensive tutorial. Explore #TidyTuesday data on San Francisco trees to build, preprocess, and optimize a predictive model. Discover techniques for data visualization, feature engineering, and model evaluation. Master parallel processing for efficient tuning, interpret results through informative plots, and implement various grid search strategies. Finalize the model specification, assess variable importance, and evaluate performance on test data. Gain practical insights into advanced machine learning workflows using R and tidymodels.
Syllabus
Introduction
Data
Species
Date variable
Map
Village Visualization
Building a model
Preprocessing
Stepother
Results
Data Preprocessing
Date Column
Downsample
Model specification
Tuning
Workflow
Preprocessor
Parallel processing
Tuning results
Plotting
Tuning parameters
Updated grid
Regular grid
Transparent grid
Finding best values
Finalizing model spec
Adding the final model
Global variable importance
Testing data
Final workflow
Last fit
Taught by
Julia Silge