Feature Engineering
The process of selecting, transforming, and creating input variables (features) from raw data to improve machine learning model performance.
Common Techniques
One-hot encoding for categorical variables. Polynomial features for non-linear relationships. Log transforms for skewed distributions. Date/time decomposition (day of week, hour, season). Text features: TF-IDF, n-grams, embeddings.
Feature Selection
Filter methods (correlation, mutual information). Wrapper methods (recursive feature elimination). Embedded methods (L1 regularization, tree-based importance). Removing redundant or irrelevant features improves both performance and training speed.
Deep Learning Impact
Deep learning reduced the need for manual feature engineering by learning representations directly from raw data. However, feature engineering remains crucial for tabular data, time series, and improving data quality before any model.