Feature engineering
Feature engineering is really an umbrella term that generally refers to two main activities: feature extraction and feature selection. Effective feature engineering can significantly enhance model performance by providing algorithms with more informative inputs and reducing or removing noisy and/or uninformative ones. These recipes will teach common approaches to feature engineering using existing features to generate new features that may (“may” being the keyword) improve model performance.
Understanding feature engineering
Feature engineering encompasses two main activities:
- Creating new features (feature extraction): This involves transforming existing data into new variables that may capture important patterns or relationships. For example, you might derive a total spending feature by combining price and quantity features.
- Selecting relevant features (feature selection): This process identifies and retains the most informative...
