1.5 The modeling process

The modeling process.

• Exploratory data analysis: Explore the data to see what they might tell you. (See previous)
• Feature engineering: Create specific model terms that make it easier to accurately model the observed data. Covered in Chapter 6.
• Model tuning and selection: Generate a variety of models and compare performance.
  • Some models require hyperparameter tuning
• Model evaluation: Use EDA-like analyses and compare model performance metrics to choose the best model for your situation.

The final model may be used for a conclusion and/or produce predictions on new data.